[X86] AVX512FP16 instructions enabling 6/6

[pengfei]

Summary

Enable FP16 complex FMA instructions.

Ref.: https://software.intel.com/content/www/us/en/develop/download/intel-avx512-fp16-architecture-specification.html

[RKSimon]

llvm/lib/Target/X86/X86ScheduleZnver3.td

64 ↗

(On Diff #355794)

You could avoid this change if you add a scheduler class to whatever instruction is complaining?

[lebedev.ri]

llvm/lib/Target/X86/X86ScheduleZnver3.td

64 ↗

(On Diff #355794)

/me goes to add herald rule that i forgot to add
Yep, please just add the new sched class here as unsupported.

[pengfei]

Add the missing scheduler info. Thanks Simon and Roman.

[pengfei]

Fix the capitalization mismatch in tests.

[LuoYuanke]

llvm/test/CodeGen/X86/avx512cfma-intrinsics.ll

4

Do we miss broadcast test case?

llvm/test/CodeGen/X86/avx512cfmul-intrinsics.ll

4

Do we miss broadcast test case?

[RKSimon]

clang/lib/Headers/avx512fp16intrin.h

2955

Outer brackets

[pengfei]

1. Rebase.
2. Add _mm_mask3_fcmadd_sch and _mm_mask3_fcmadd_round_sch.
3. Address comments from Yuanke and Simon.

[LuoYuanke]

clang/test/CodeGen/X86/avx512fp16-builtins.c

4223

MADD?

4315

MADD?

llvm/include/llvm/IR/IntrinsicsX86.td

5736

_cph?

5800

_csh?

llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp

3903

"b" means rounding. Right?

3949

Sorry, I didn't find the constrain in the spec.

llvm/lib/Target/X86/X86ISelLowering.cpp

47414

Can swap LHS and RHS reduce some redundant code?

47421

The lambda seems only be called once.

47423

Is it possible fast and non-fast instruction is mixed due to inline? Shall we check the instruction AllowContract flag?

47436

Merge it to previous line.

llvm/lib/Target/X86/X86InstrAVX512.td

5772

Moving ClobberConstraint before IsCommutable saves the code for default value?

13593

The name seems not accurate. Is it cfmop for mul and cfmaop for fma?

13629

I didn't see this flag for other scalar instructions, why we need it for complex instruction?

llvm/lib/Target/X86/X86InstrFoldTables.cpp

1852

Why FR32X version is not needed for complex scalar instructions?

[pengfei]

llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp

3903

broadcasting

3949

#UD if (dest_reg == src1_reg) or ( dest_reg == src2_reg)

llvm/lib/Target/X86/X86InstrAVX512.td

13629

Because all complex instructions have constrains "dst != src1 && dst != src2". We use earlyclobber to avoid the dst been assigned to src1 or src2.

[pengfei]

llvm/lib/Target/X86/X86InstrFoldTables.cpp

1852

Do you mean complex ss/sd? We don't have these instructions.

[LuoYuanke]

llvm/lib/Target/X86/X86InstrFoldTables.cpp

1852

No, I mean we have both X86::XXX and X86::XXX_Int for other instructions. One is FR16X which can be unfolded, one is VR128X which can't. For example, VFNMADD213SHZm and VFNMADD213SHZm_Int.

[craig.topper]

llvm/lib/Target/X86/X86InstrFoldTables.cpp

1852

The VFCMULCSHZrr instructions produce two 16-bit values packed into the lower 32 bits. That would mean we would need a FR32X result, but it couldn't interact meaningfully with any other FR32X instruction since its really two values.

I think we only have FR32/FR64 instructions for things that have generic IR equivalents or that we create from other generic IR operations. Like I think we have an FR32 RCP and RSQRT because we can convert float div or 1/sqrt to them.

[LuoYuanke]

llvm/lib/Target/X86/X86InstrFoldTables.cpp

1852

Thanks, Craig. I understand now. :)

[LuoYuanke]

llvm/lib/Target/X86/X86InstrAVX512.td

13629

Got it. Thanks!

[pengfei]

Address Yuanke's comments.

[pengfei]

Thanks for the review.

clang/test/CodeGen/X86/avx512fp16-builtins.c

4223

They are marks used when adding tests. We can remove them now.

[LuoYuanke]

llvm/lib/Target/X86/X86ISelLowering.cpp

47544

Sorry, I don't understand the comments. What does FMF mean?

[RKSimon]

llvm/lib/Target/X86/X86ISelLowering.cpp

47544

fast math flags?

[LuoYuanke]

llvm/lib/Target/X86/X86ISelLowering.cpp

47544

I understand now. Thanks, Simon. :)

[pengfei]

Update comments. NFC.

llvm/lib/Target/X86/X86ISelLowering.cpp

47544

Yes.

[LuoYuanke]

LGTM, but may wait 1 or 2 days for the comments from others.

[pengfei]

llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp

3903

Sorry, my mistake. Here b supposes to represent EVEX.b bit in the encoding.
It's used as broadcast control only in memory variants in vector instructions.
When it is used in register variants, it enables rounding control and SAE.

clang/include/clang/Basic/BuiltinsX86.def

 TARGET_BUILTIN(__builtin_ia32_vfmsubaddph512_mask3, "V32xV32xV32xV32xUiIi", "ncV:512:", "avx512fp16")
 TARGET_BUILTIN(__builtin_ia32_vfmsubph512_mask3, "V32xV32xV32xV32xUiIi", "ncV:512:", "avx512fp16")
 
 TARGET_BUILTIN(__builtin_ia32_vfmaddsh3_mask, "V8xV8xV8xV8xUcIi", "ncV:128:", "avx512fp16")
 TARGET_BUILTIN(__builtin_ia32_vfmaddsh3_maskz, "V8xV8xV8xV8xUcIi", "ncV:128:", "avx512fp16")
 TARGET_BUILTIN(__builtin_ia32_vfmaddsh3_mask3, "V8xV8xV8xV8xUcIi", "ncV:128:", "avx512fp16")
 TARGET_BUILTIN(__builtin_ia32_vfmsubsh3_mask3, "V8xV8xV8xV8xUcIi", "ncV:128:", "avx512fp16")
 
+TARGET_BUILTIN(__builtin_ia32_vfmaddcph128_mask,  "V4fV4fV4fV4fUc", "ncV:128:", "avx512fp16,avx512vl")
+TARGET_BUILTIN(__builtin_ia32_vfmaddcph128_maskz,  "V4fV4fV4fV4fUc", "ncV:128:", "avx512fp16,avx512vl")
+TARGET_BUILTIN(__builtin_ia32_vfmaddcph256_mask,  "V8fV8fV8fV8fUc", "ncV:256:", "avx512fp16,avx512vl")
+TARGET_BUILTIN(__builtin_ia32_vfmaddcph256_maskz,  "V8fV8fV8fV8fUc", "ncV:256:", "avx512fp16,avx512vl")
+TARGET_BUILTIN(__builtin_ia32_vfmaddcph512_mask,  "V16fV16fV16fV16fUsIi", "ncV:512:", "avx512fp16")
+TARGET_BUILTIN(__builtin_ia32_vfmaddcph512_maskz,  "V16fV16fV16fV16fUsIi", "ncV:512:", "avx512fp16")
+TARGET_BUILTIN(__builtin_ia32_vfcmaddcph128_mask,  "V4fV4fV4fV4fUc", "ncV:128:", "avx512fp16,avx512vl")
+TARGET_BUILTIN(__builtin_ia32_vfcmaddcph128_maskz,  "V4fV4fV4fV4fUc", "ncV:128:", "avx512fp16,avx512vl")
+TARGET_BUILTIN(__builtin_ia32_vfcmaddcph256_mask,  "V8fV8fV8fV8fUc", "ncV:256:", "avx512fp16,avx512vl")
+TARGET_BUILTIN(__builtin_ia32_vfcmaddcph256_maskz,  "V8fV8fV8fV8fUc", "ncV:256:", "avx512fp16,avx512vl")
+TARGET_BUILTIN(__builtin_ia32_vfcmaddcph512_mask,  "V16fV16fV16fV16fUsIi", "ncV:512:", "avx512fp16")
+TARGET_BUILTIN(__builtin_ia32_vfcmaddcph512_maskz,  "V16fV16fV16fV16fUsIi", "ncV:512:", "avx512fp16")
+TARGET_BUILTIN(__builtin_ia32_vfmaddcsh_mask,   "V4fV4fV4fV4fUcIi", "ncV:128:", "avx512fp16")
+TARGET_BUILTIN(__builtin_ia32_vfmaddcsh_maskz,   "V4fV4fV4fV4fUcIi", "ncV:128:", "avx512fp16")
+TARGET_BUILTIN(__builtin_ia32_vfcmaddcsh_mask,  "V4fV4fV4fV4fUcIi", "ncV:128:", "avx512fp16")
+TARGET_BUILTIN(__builtin_ia32_vfcmaddcsh_maskz,  "V4fV4fV4fV4fUcIi", "ncV:128:", "avx512fp16")
+
+TARGET_BUILTIN(__builtin_ia32_vfmulcsh_mask,   "V4fV4fV4fV4fUcIi", "ncV:128:", "avx512fp16")
+TARGET_BUILTIN(__builtin_ia32_vfcmulcsh_mask,  "V4fV4fV4fV4fUcIi", "ncV:128:", "avx512fp16")
+TARGET_BUILTIN(__builtin_ia32_vfmulcph128_mask,  "V4fV4fV4fV4fUc", "ncV:128:", "avx512fp16,avx512vl")
+TARGET_BUILTIN(__builtin_ia32_vfmulcph256_mask,  "V8fV8fV8fV8fUc", "ncV:256:", "avx512fp16,avx512vl")
+TARGET_BUILTIN(__builtin_ia32_vfmulcph512_mask,  "V16fV16fV16fV16fUsIi", "ncV:512:", "avx512fp16")
+TARGET_BUILTIN(__builtin_ia32_vfcmulcph128_mask,  "V4fV4fV4fV4fUc", "ncV:128:", "avx512fp16,avx512vl")
+TARGET_BUILTIN(__builtin_ia32_vfcmulcph256_mask,  "V8fV8fV8fV8fUc", "ncV:256:", "avx512fp16,avx512vl")
+TARGET_BUILTIN(__builtin_ia32_vfcmulcph512_mask,  "V16fV16fV16fV16fUsIi", "ncV:512:", "avx512fp16")
+
 // generic select intrinsics
 TARGET_BUILTIN(__builtin_ia32_selectb_128, "V16cUsV16cV16c", "ncV:128:", "avx512bw,avx512vl")
 TARGET_BUILTIN(__builtin_ia32_selectb_256, "V32cUiV32cV32c", "ncV:256:", "avx512bw,avx512vl")
 TARGET_BUILTIN(__builtin_ia32_selectb_512, "V64cUOiV64cV64c", "ncV:512:", "avx512bw")
 TARGET_BUILTIN(__builtin_ia32_selectw_128, "V8sUcV8sV8s", "ncV:128:", "avx512bw,avx512vl")
 TARGET_BUILTIN(__builtin_ia32_selectw_256, "V16sUsV16sV16s", "ncV:256:", "avx512bw,avx512vl")
 TARGET_BUILTIN(__builtin_ia32_selectw_512, "V32sUiV32sV32s", "ncV:512:", "avx512bw")
 TARGET_BUILTIN(__builtin_ia32_selectd_128, "V4iUcV4iV4i", "ncV:128:", "avx512vl")

clang/lib/Headers/avx512fp16intrin.h

#define _mm512_maskz_max_round_ph(U, A, B, R)                                  \
   ((__m512h)__builtin_ia32_selectph_512(                                       \
       (__mmask32)(U), (__v32hf)_mm512_max_round_ph((A), (B), (R)),             \
       (__v32hf)_mm512_setzero_ph()))
 
 static __inline__ __m512h __DEFAULT_FN_ATTRS512 _mm512_abs_ph(__m512h __A) {
   return (__m512h)_mm512_and_epi32(_mm512_set1_epi32(0x7FFF7FFF), (__m512i)__A);
 }
 
+static __inline__ __m512h __DEFAULT_FN_ATTRS512 _mm512_conj_pch(__m512h __A) {
+  return (__m512h)_mm512_xor_ps((__m512)__A, _mm512_set1_ps(-0.0f));
+}
+
+static __inline__ __m512h __DEFAULT_FN_ATTRS512
+_mm512_mask_conj_pch(__m512h __W, __mmask16 __U, __m512h __A) {
+  return (__m512h)__builtin_ia32_selectps_512(
+      (__mmask16)__U, (__v16sf)_mm512_conj_pch(__A), (__v16sf)__W);
+}
+
+static __inline__ __m512h __DEFAULT_FN_ATTRS512
+_mm512_maskz_conj_pch(__mmask16 __U, __m512h __A) {
+  return (__m512h)__builtin_ia32_selectps_512((__mmask16)__U,
+                                              (__v16sf)_mm512_conj_pch(__A),
+                                              (__v16sf)_mm512_setzero_ps());
+}
+
 static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_add_sh(__m128h __A,
                                                            __m128h __B) {
   __A[0] += __B[0];
   return __A;
 }
 
 static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_mask_add_sh(__m128h __W,
                                                                 __mmask8 __U,
return __builtin_ia32_vfmsubsh3_mask3((__v8hf)__W, -(__v8hf)__X, (__v8hf)__Y,
                                         _MM_FROUND_CUR_DIRECTION);
 }
 
 #define _mm_mask3_fnmsub_round_sh(W, X, Y, U, R)                               \
   ((__m128h)__builtin_ia32_vfmsubsh3_mask3(                                    \
       (__v8hf)(__m128h)(W), -(__v8hf)(__m128h)(X), (__v8hf)(__m128h)(Y),       \
       (__mmask8)(U), (int)(R)))
 
+static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_fcmadd_sch(__m128h __A,
+                                                               __m128h __B,
+                                                               __m128h __C) {
+  return (__m128h)__builtin_ia32_vfcmaddcsh_mask((__v4sf)__C, (__v4sf)__A,
+                                                 (__v4sf)__B, (__mmask8)-1,
+                                                 _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_mask_fcmadd_sch(__m128h __A, __mmask8 __U, __m128h __B, __m128h __C) {
+  return (__m128h)__builtin_ia32_selectps_128(
+      __U,
+      __builtin_ia32_vfcmaddcsh_mask((__v4sf)__C, (__v4sf)__A, (__v4sf)__B,
+                                     (__mmask8)__U, _MM_FROUND_CUR_DIRECTION),
+      (__v4sf)__A);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_maskz_fcmadd_sch(__mmask8 __U, __m128h __A, __m128h __B, __m128h __C) {
+  return (__m128h)__builtin_ia32_vfcmaddcsh_maskz((__v4sf)__C, (__v4sf)__A,
+                                                  (__v4sf)__B, (__mmask8)__U,
+                                                  _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_mask3_fcmadd_sch(__m128h __A, __m128h __B, __m128h __C, __mmask8 __U) {
+  return (__m128h)_mm_move_ss((__m128)__C,
+                              (__m128)__builtin_ia32_vfcmaddcsh_mask(
+                                  (__v4sf)__C, (__v4sf)__A, (__v4sf)__B, __U,
+                                  _MM_FROUND_CUR_DIRECTION));
+}
+
+#define _mm_fcmadd_round_sch(A, B, C, R)                                       \
+  ((__m128h)__builtin_ia32_vfcmaddcsh_mask(                                    \
+      (__v4sf)(__m128h)(C), (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B),        \
+      (__mmask8)-1, (int)(R)))
+
+#define _mm_mask_fcmadd_round_sch(A, U, B, C, R)                               \
+  ((__m128h)__builtin_ia32_selectps_128(                                       \
+      (__mmask8)(U & 1),                                                       \
+      __builtin_ia32_vfcmaddcsh_mask(                                          \
+          (__v4sf)(__m128h)(C), (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B),    \
+          (__mmask8)(U), (int)(R)),                                            \
+      (__v4sf)(__m128h)(A)))
+
+#define _mm_maskz_fcmadd_round_sch(U, A, B, C, R)                              \
+  ((__m128h)__builtin_ia32_vfcmaddcsh_maskz(                                   \
+      (__v4sf)(__m128h)(C), (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B),        \
+      (__mmask8)(U), (int)(R)))
+
+#define _mm_mask3_fcmadd_round_sch(A, B, C, U, R)                              \
+  ((__m128h)_mm_move_ss((__m128)(C),                                           \
+                        (__m128)__builtin_ia32_vfcmaddcsh_mask(                \
+                            (__v4sf)(C), (__v4sf)(A), (__v4sf)(B), (U), (R))))
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_fmadd_sch(__m128h __A,
+                                                              __m128h __B,
+                                                              __m128h __C) {
+  return (__m128h)__builtin_ia32_vfmaddcsh_mask((__v4sf)__C, (__v4sf)__A,
+                                                (__v4sf)__B, (__mmask8)-1,
+                                                _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_mask_fmadd_sch(__m128h __A, __mmask8 __U, __m128h __B, __m128h __C) {
+  return (__m128h)__builtin_ia32_selectps_128(
+      __U,
+      __builtin_ia32_vfmaddcsh_mask((__v4sf)__C, (__v4sf)__A, (__v4sf)__B,
+                                    (__mmask8)__U, _MM_FROUND_CUR_DIRECTION),
+      (__v4sf)__A);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_maskz_fmadd_sch(__mmask8 __U, __m128h __A, __m128h __B, __m128h __C) {
+  return (__m128h)__builtin_ia32_vfmaddcsh_maskz((__v4sf)__C, (__v4sf)__A,
+                                                 (__v4sf)__B, (__mmask8)__U,
+                                                 _MM_FROUND_CUR_DIRECTION);
+}
+
+#define _mm_fmadd_round_sch(A, B, C, R)                                        \
+  ((__m128h)__builtin_ia32_vfmaddcsh_mask(                                     \
+      (__v4sf)(__m128h)(C), (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B),        \
+      (__mmask8)-1, (int)(R)))
+
+#define _mm_mask_fmadd_round_sch(A, U, B, C, R)                                \
+  ((__m128h)__builtin_ia32_selectps_128(                                       \
+      (__mmask8)(U & 1),                                                       \
+      __builtin_ia32_vfmaddcsh_mask(                                           \
+          (__v4sf)(__m128h)(C), (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B),    \
+          (__mmask8)(U), (int)(R)),                                            \
+      (__v4sf)(__m128h)(A)))
+
+#define _mm_maskz_fmadd_round_sch(U, A, B, C, R)                               \
+  ((__m128h)__builtin_ia32_vfmaddcsh_maskz(                                    \
+      (__v4sf)(__m128h)(C), (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B),        \
+      (__mmask8)(U), (int)(R)))
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_fcmul_sch(__m128h __A,
+                                                              __m128h __B) {
+  return (__m128h)__builtin_ia32_vfcmulcsh_mask(
+      (__v4sf)__A, (__v4sf)__B, (__v4sf)_mm_undefined_ph(), (__mmask8)-1,
+      _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_mask_fcmul_sch(__m128h __W, __mmask8 __U, __m128h __A, __m128h __B) {
+  return (__m128h)__builtin_ia32_vfcmulcsh_mask((__v4sf)__A, (__v4sf)__B,
+                                                (__v4sf)__W, (__mmask8)__U,
+                                                _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_maskz_fcmul_sch(__mmask8 __U, __m128h __A, __m128h __B) {
+  return (__m128h)__builtin_ia32_vfcmulcsh_mask(
+      (__v4sf)__A, (__v4sf)__B, (__v4sf)_mm_setzero_ph(), (__mmask8)__U,
+      _MM_FROUND_CUR_DIRECTION);
+}
+
+#define _mm_fcmul_round_sch(A, B, R)                                           \
+  ((__m128h)__builtin_ia32_vfcmulcsh_mask(                                     \
+      (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B),                              \
+      (__v4sf)(__m128h)_mm_undefined_ph(), (__mmask8)-1, (int)(R)))
+
+#define _mm_mask_fcmul_round_sch(W, U, A, B, R)                                \
+  ((__m128h)__builtin_ia32_vfcmulcsh_mask(                                     \
+      (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B), (__v4sf)(__m128h)(W),        \
+      (__mmask8)(U), (int)(R)))
+
+#define _mm_maskz_fcmul_round_sch(U, A, B, R)                                  \
+  ((__m128h)__builtin_ia32_vfcmulcsh_mask(                                     \
+      (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B),                              \
+      (__v4sf)(__m128h)_mm_setzero_ph(), (__mmask8)(U), (int)(R)))
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_fmul_sch(__m128h __A,
+                                                             __m128h __B) {
+  return (__m128h)__builtin_ia32_vfmulcsh_mask(
+      (__v4sf)__A, (__v4sf)__B, (__v4sf)_mm_undefined_ph(), (__mmask8)-1,
+      _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_mask_fmul_sch(__m128h __W,
+                                                                  __mmask8 __U,
+                                                                  __m128h __A,
+                                                                  __m128h __B) {
+  return (__m128h)__builtin_ia32_vfmulcsh_mask((__v4sf)__A, (__v4sf)__B,
+                                               (__v4sf)__W, (__mmask8)__U,
+                                               _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_maskz_fmul_sch(__mmask8 __U, __m128h __A, __m128h __B) {
+  return (__m128h)__builtin_ia32_vfmulcsh_mask(
+      (__v4sf)__A, (__v4sf)__B, (__v4sf)_mm_setzero_ph(), (__mmask8)__U,
+      _MM_FROUND_CUR_DIRECTION);
+}
+
+#define _mm_fmul_round_sch(A, B, R)                                            \
+  ((__m128h)__builtin_ia32_vfmulcsh_mask(                                      \
+      (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B),                              \
+      (__v4sf)(__m128h)_mm_undefined_ph(), (__mmask8)-1, (int)(R)))
+
+#define _mm_mask_fmul_round_sch(W, U, A, B, R)                                 \
+  ((__m128h)__builtin_ia32_vfmulcsh_mask(                                      \
+      (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B), (__v4sf)(__m128h)(W),        \
+      (__mmask8)(U), (int)(R)))
+
+#define _mm_maskz_fmul_round_sch(U, A, B, R)                                   \
+  ((__m128h)__builtin_ia32_vfmulcsh_mask(                                      \
+      (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B),                              \
+      (__v4sf)(__m128h)_mm_setzero_ph(), (__mmask8)(U), (int)(R)))
+
+static __inline__ __m512h __DEFAULT_FN_ATTRS512 _mm512_fcmul_pch(__m512h __A,
+                                                                 __m512h __B) {
+  return (__m512h)__builtin_ia32_vfcmulcph512_mask(
+      (__v16sf)__A, (__v16sf)__B, (__v16sf)_mm512_undefined_ph(), (__mmask16)-1,
+      _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m512h __DEFAULT_FN_ATTRS512
+_mm512_mask_fcmul_pch(__m512h __W, __mmask16 __U, __m512h __A, __m512h __B) {
+  return (__m512h)__builtin_ia32_vfcmulcph512_mask((__v16sf)__A, (__v16sf)__B,
+                                                   (__v16sf)__W, (__mmask16)__U,
+                                                   _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m512h __DEFAULT_FN_ATTRS512
+_mm512_maskz_fcmul_pch(__mmask16 __U, __m512h __A, __m512h __B) {
+  return (__m512h)__builtin_ia32_vfcmulcph512_mask(
+      (__v16sf)__A, (__v16sf)__B, (__v16sf)_mm512_setzero_ph(), (__mmask16)__U,
+      _MM_FROUND_CUR_DIRECTION);
+}
+
+#define _mm512_fcmul_round_pch(A, B, R)                                        \
+  ((__m512h)__builtin_ia32_vfcmulcph512_mask(                                  \
+      (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B),                            \
+      (__v16sf)(__m512h)_mm512_undefined_ph(), (__mmask16)-1, (int)(R)))
+
+#define _mm512_mask_fcmul_round_pch(W, U, A, B, R)                             \
+  ((__m512h)__builtin_ia32_vfcmulcph512_mask(                                  \
+      (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B), (__v16sf)(__m512h)(W),     \
+      (__mmask16)(U), (int)(R)))
+
+#define _mm512_maskz_fcmul_round_pch(U, A, B, R)                               \
+  ((__m512h)__builtin_ia32_vfcmulcph512_mask(                                  \
+      (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B),                            \
+      (__v16sf)(__m512h)_mm512_setzero_ph(), (__mmask16)(U), (int)(R)))
+
+static __inline__ __m512h __DEFAULT_FN_ATTRS512 _mm512_fmul_pch(__m512h __A,
+                                                                __m512h __B) {
+  return (__m512h)__builtin_ia32_vfmulcph512_mask(
+      (__v16sf)__A, (__v16sf)__B, (__v16sf)_mm512_undefined_ph(), (__mmask16)-1,
+      _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m512h __DEFAULT_FN_ATTRS512
+_mm512_mask_fmul_pch(__m512h __W, __mmask16 __U, __m512h __A, __m512h __B) {
+  return (__m512h)__builtin_ia32_vfmulcph512_mask((__v16sf)__A, (__v16sf)__B,
+                                                  (__v16sf)__W, (__mmask16)__U,
+                                                  _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m512h __DEFAULT_FN_ATTRS512
+_mm512_maskz_fmul_pch(__mmask16 __U, __m512h __A, __m512h __B) {
+  return (__m512h)__builtin_ia32_vfmulcph512_mask(
+      (__v16sf)__A, (__v16sf)__B, (__v16sf)_mm512_setzero_ph(), (__mmask16)__U,
+      _MM_FROUND_CUR_DIRECTION);
+}
+
+#define _mm512_fmul_round_pch(A, B, R)                                         \
+  ((__m512h)__builtin_ia32_vfmulcph512_mask(                                   \
+      (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B),                            \
+      (__v16sf)(__m512h)_mm512_undefined_ph(), (__mmask16)-1, (int)(R)))
+
+#define _mm512_mask_fmul_round_pch(W, U, A, B, R)                              \
+  ((__m512h)__builtin_ia32_vfmulcph512_mask(                                   \
+      (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B), (__v16sf)(__m512h)(W),     \
+      (__mmask16)(U), (int)(R)))
+
+#define _mm512_maskz_fmul_round_pch(U, A, B, R)                                \
+  ((__m512h)__builtin_ia32_vfmulcph512_mask(                                   \
+      (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B),                            \
+      (__v16sf)(__m512h)_mm512_setzero_ph(), (__mmask16)(U), (int)(R)))
+
+static __inline__ __m512h __DEFAULT_FN_ATTRS512 _mm512_fcmadd_pch(__m512h __A,
+                                                                  __m512h __B,
+                                                                  __m512h __C) {
+  return (__m512h)__builtin_ia32_vfcmaddcph512_mask((__v16sf)__C, (__v16sf)__A,
+                                                    (__v16sf)__B, (__mmask16)-1,
+                                                    _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m512h __DEFAULT_FN_ATTRS512
+_mm512_mask_fcmadd_pch(__m512h __A, __mmask16 __U, __m512h __B, __m512h __C) {
+  return (__m512h)__builtin_ia32_selectps_512(
+      __U,
+      __builtin_ia32_vfcmaddcph512_mask((__v16sf)__C, (__v16sf)__A,
+                                        (__v16sf)__B, (__mmask16)__U,
+                                        _MM_FROUND_CUR_DIRECTION),
+      (__v16sf)__A);
+}
+
+static __inline__ __m512h __DEFAULT_FN_ATTRS512
+_mm512_mask3_fcmadd_pch(__m512h __A, __m512h __B, __m512h __C, __mmask16 __U) {
+  return (__m512h)__builtin_ia32_vfcmaddcph512_mask(
+      (__v16sf)__C, (__v16sf)__A, (__v16sf)__B, (__mmask16)__U,
+      _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m512h __DEFAULT_FN_ATTRS512
+_mm512_maskz_fcmadd_pch(__mmask16 __U, __m512h __A, __m512h __B, __m512h __C) {
+  return (__m512h)__builtin_ia32_vfcmaddcph512_maskz(
+      (__v16sf)__C, (__v16sf)__A, (__v16sf)__B, (__mmask16)__U,
+      _MM_FROUND_CUR_DIRECTION);
+}
+
+#define _mm512_fcmadd_round_pch(A, B, C, R)                                    \
+  ((__m512h)__builtin_ia32_vfcmaddcph512_mask(                                 \
+      (__v16sf)(__m512h)(C), (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B),     \
+      (__mmask16)-1, (int)(R)))
+
+#define _mm512_mask_fcmadd_round_pch(A, U, B, C, R)                            \
+  ((__m512h)__builtin_ia32_selectps_512(                                       \
+      (__mmask16)(U),                                                          \
+      __builtin_ia32_vfcmaddcph512_mask(                                       \
+          (__v16sf)(__m512h)(C), (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B), \
+          (__mmask16)(U), (int)(R)),                                           \
+      (__v16sf)(__m512h)(A)))
+
+#define _mm512_mask3_fcmadd_round_pch(A, B, C, U, R)                           \
+  ((__m512h)__builtin_ia32_vfcmaddcph512_mask(                                 \
+      (__v16sf)(__m512h)(C), (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B),     \
+      (__mmask16)(U), (int)(R)))
+
+#define _mm512_maskz_fcmadd_round_pch(U, A, B, C, R)                           \
+  ((__m512h)__builtin_ia32_vfcmaddcph512_maskz(                                \
+      (__v16sf)(__m512h)(C), (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B),     \
+      (__mmask16)(U), (int)(R)))
+
+static __inline__ __m512h __DEFAULT_FN_ATTRS512 _mm512_fmadd_pch(__m512h __A,
+                                                                 __m512h __B,
+                                                                 __m512h __C) {
+  return (__m512h)__builtin_ia32_vfmaddcph512_mask((__v16sf)__C, (__v16sf)__A,
+                                                   (__v16sf)__B, (__mmask16)-1,
+                                                   _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m512h __DEFAULT_FN_ATTRS512
+_mm512_mask_fmadd_pch(__m512h __A, __mmask16 __U, __m512h __B, __m512h __C) {
+  return (__m512h)__builtin_ia32_selectps_512(
+      __U,
+      __builtin_ia32_vfmaddcph512_mask((__v16sf)__C, (__v16sf)__A, (__v16sf)__B,
+                                       (__mmask16)__U,
+                                       _MM_FROUND_CUR_DIRECTION),
+      (__v16sf)__A);
+}
+
+static __inline__ __m512h __DEFAULT_FN_ATTRS512
+_mm512_mask3_fmadd_pch(__m512h __A, __m512h __B, __m512h __C, __mmask16 __U) {
+  return (__m512h)__builtin_ia32_vfmaddcph512_mask((__v16sf)__C, (__v16sf)__A,
+                                                   (__v16sf)__B, (__mmask16)__U,
+                                                   _MM_FROUND_CUR_DIRECTION);
+}
+
+static __inline__ __m512h __DEFAULT_FN_ATTRS512
+_mm512_maskz_fmadd_pch(__mmask16 __U, __m512h __A, __m512h __B, __m512h __C) {
+  return (__m512h)__builtin_ia32_vfmaddcph512_maskz(
+      (__v16sf)__C, (__v16sf)__A, (__v16sf)__B, (__mmask16)__U,
+      _MM_FROUND_CUR_DIRECTION);
+}
+
+#define _mm512_fmadd_round_pch(A, B, C, R)                                     \
+  ((__m512h)__builtin_ia32_vfmaddcph512_mask(                                  \
+      (__v16sf)(__m512h)(C), (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B),     \
+      (__mmask16)-1, (int)(R)))
+
+#define _mm512_mask_fmadd_round_pch(A, U, B, C, R)                             \
+  ((__m512h)__builtin_ia32_selectps_512(                                       \
+      (__mmask16)(U),                                                          \
+      __builtin_ia32_vfmaddcph512_mask(                                        \
+          (__v16sf)(__m512h)(C), (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B), \
+          (__mmask16)(U), (int)(R)),                                           \
+      (__v16sf)(__m512h)(A)))
+
+#define _mm512_mask3_fmadd_round_pch(A, B, C, U, R)                            \
+  ((__m512h)__builtin_ia32_vfmaddcph512_mask(                                  \
+      (__v16sf)(__m512h)(C), (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B),     \
+      (__mmask16)(U), (int)(R)))
+
+#define _mm512_maskz_fmadd_round_pch(U, A, B, C, R)                            \
+  ((__m512h)__builtin_ia32_vfmaddcph512_maskz(                                 \
+      (__v16sf)(__m512h)(C), (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B),     \
+      (__mmask16)(U), (int)(R)))
+
 static __inline__ _Float16 __DEFAULT_FN_ATTRS512
 _mm512_reduce_add_ph(__m512h __W) {
   return __builtin_ia32_reduce_fadd_ph512(-0.0f16, __W);
 }
 
 static __inline__ _Float16 __DEFAULT_FN_ATTRS512
 _mm512_reduce_mul_ph(__m512h __W) {
   return __builtin_ia32_reduce_fmul_ph512(1.0f16, __W);

clang/lib/Headers/avx512vlfp16intrin.h

 static __inline__ __m256h __DEFAULT_FN_ATTRS256 _mm256_abs_ph(__m256h __A) {
   return (__m256h)_mm256_and_epi32(_mm256_set1_epi32(0x7FFF7FFF), (__m256i)__A);
 }
 
 static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_abs_ph(__m128h __A) {
   return (__m128h)_mm_and_epi32(_mm_set1_epi32(0x7FFF7FFF), (__m128i)__A);
 }
 
+static __inline__ __m256h __DEFAULT_FN_ATTRS256 _mm256_conj_pch(__m256h __A) {
+  return (__m256h)_mm256_xor_ps((__m256)__A, _mm256_set1_ps(-0.0f));
+}
+
+static __inline__ __m256h __DEFAULT_FN_ATTRS256
+_mm256_mask_conj_pch(__m256h __W, __mmask8 __U, __m256h __A) {
+  return (__m256h)__builtin_ia32_selectps_256(
+      (__mmask8)__U, (__v8sf)_mm256_conj_pch(__A), (__v8sf)__W);
+}
+
+static __inline__ __m256h __DEFAULT_FN_ATTRS256
+_mm256_maskz_conj_pch(__mmask8 __U, __m256h __A) {
+  return (__m256h)__builtin_ia32_selectps_256(
+      (__mmask8)__U, (__v8sf)_mm256_conj_pch(__A), (__v8sf)_mm256_setzero_ps());
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_conj_pch(__m128h __A) {
+  return (__m128h)_mm_xor_ps((__m128)__A, _mm_set1_ps(-0.0f));
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_mask_conj_pch(__m128h __W,
+                                                                  __mmask8 __U,
+                                                                  __m128h __A) {
+  return (__m128h)__builtin_ia32_selectps_128(
+      (__mmask8)__U, (__v4sf)_mm_conj_pch(__A), (__v4sf)__W);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_maskz_conj_pch(__mmask8 __U, __m128h __A) {
+  return (__m128h)__builtin_ia32_selectps_128(
+      (__mmask8)__U, (__v4sf)_mm_conj_pch(__A), (__v4sf)_mm_setzero_ps());
+}
+
 #define _mm256_cmp_ph_mask(a, b, p)                                            \
   ((__mmask16)__builtin_ia32_cmpph256_mask(                                    \
       (__v16hf)(__m256h)(a), (__v16hf)(__m256h)(b), (int)(p), (__mmask16)-1))
 
 #define _mm256_mask_cmp_ph_mask(m, a, b, p)                                    \
   ((__mmask16)__builtin_ia32_cmpph256_mask(                                    \
       (__v16hf)(__m256h)(a), (__v16hf)(__m256h)(b), (int)(p), (__mmask16)(m)))
 

 static __inline__ __m256h __DEFAULT_FN_ATTRS256
 _mm256_mask3_fnmsub_ph(__m256h __A, __m256h __B, __m256h __C, __mmask16 __U) {
   return (__m256h)__builtin_ia32_selectph_256(
       (__mmask16)__U,
       __builtin_ia32_vfmaddph256((__v16hf)__A, -(__v16hf)__B, -(__v16hf)__C),
       (__v16hf)__C);
 }
 
+static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_fcmul_pch(__m128h __A,
+                                                              __m128h __B) {
+  return (__m128h)__builtin_ia32_vfcmulcph128_mask(
+      (__v4sf)__A, (__v4sf)__B, (__v4sf)_mm_undefined_ph(), (__mmask8)-1);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_mask_fcmul_pch(__m128h __W, __mmask8 __U, __m128h __A, __m128h __B) {
+  return (__m128h)__builtin_ia32_vfcmulcph128_mask((__v4sf)__A, (__v4sf)__B,
+                                                   (__v4sf)__W, (__mmask8)__U);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_maskz_fcmul_pch(__mmask8 __U, __m128h __A, __m128h __B) {
+  return (__m128h)__builtin_ia32_vfcmulcph128_mask(
+      (__v4sf)__A, (__v4sf)__B, (__v4sf)_mm_setzero_ph(), (__mmask8)__U);
+}
+
+static __inline__ __m256h __DEFAULT_FN_ATTRS128 _mm256_fcmul_pch(__m256h __A,
+                                                                 __m256h __B) {
+  return (__m256h)__builtin_ia32_vfcmulcph256_mask(
+      (__v8sf)__A, (__v8sf)__B, (__v8sf)_mm256_undefined_ph(), (__mmask8)-1);
+}
+
+static __inline__ __m256h __DEFAULT_FN_ATTRS256
+_mm256_mask_fcmul_pch(__m256h __W, __mmask8 __U, __m256h __A, __m256h __B) {
+  return (__m256h)__builtin_ia32_vfcmulcph256_mask((__v8sf)__A, (__v8sf)__B,
+                                                   (__v8sf)__W, (__mmask8)__U);
+}
+
+static __inline__ __m256h __DEFAULT_FN_ATTRS256
+_mm256_maskz_fcmul_pch(__mmask8 __U, __m256h __A, __m256h __B) {
+  return (__m256h)__builtin_ia32_vfcmulcph256_mask(
+      (__v8sf)__A, (__v8sf)__B, (__v8sf)_mm256_setzero_ph(), (__mmask8)__U);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_fcmadd_pch(__m128h __A,
+                                                               __m128h __B,
+                                                               __m128h __C) {
+  return (__m128h)__builtin_ia32_vfcmaddcph128_mask((__v4sf)__C, (__v4sf)__A,
+                                                    (__v4sf)__B, (__mmask8)-1);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_mask_fcmadd_pch(__m128h __A, __mmask8 __U, __m128h __B, __m128h __C) {
+  return (__m128h)__builtin_ia32_selectps_128(
+      __U,
+      __builtin_ia32_vfcmaddcph128_mask((__v4sf)__C, (__v4sf)(__m128h)__A,
+                                        (__v4sf)__B, (__mmask8)__U),
+      (__v4sf)__A);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_mask3_fcmadd_pch(__m128h __A, __m128h __B, __m128h __C, __mmask8 __U) {
+  return (__m128h)__builtin_ia32_vfcmaddcph128_mask((__v4sf)__C, (__v4sf)__A,
+                                                    (__v4sf)__B, (__mmask8)__U);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_maskz_fcmadd_pch(__mmask8 __U, __m128h __A, __m128h __B, __m128h __C) {
+  return (__m128h)__builtin_ia32_vfcmaddcph128_maskz(
+      (__v4sf)__C, (__v4sf)__A, (__v4sf)__B, (__mmask8)__U);
+}
+
+static __inline__ __m256h __DEFAULT_FN_ATTRS256 _mm256_fcmadd_pch(__m256h __A,
+                                                                  __m256h __B,
+                                                                  __m256h __C) {
+  return (__m256h)__builtin_ia32_vfcmaddcph256_mask((__v8sf)__C, (__v8sf)__A,
+                                                    (__v8sf)__B, (__mmask8)-1);
+}
+
+static __inline__ __m256h __DEFAULT_FN_ATTRS256
+_mm256_mask_fcmadd_pch(__m256h __A, __mmask8 __U, __m256h __B, __m256h __C) {
+  return (__m256h)__builtin_ia32_selectps_256(
+      __U,
+      __builtin_ia32_vfcmaddcph256_mask((__v8sf)__C, (__v8sf)__A, (__v8sf)__B,
+                                        (__mmask8)__U),
+      (__v8sf)__A);
+}
+
+static __inline__ __m256h __DEFAULT_FN_ATTRS256
+_mm256_mask3_fcmadd_pch(__m256h __A, __m256h __B, __m256h __C, __mmask8 __U) {
+  return (__m256h)__builtin_ia32_vfcmaddcph256_mask((__v8sf)__C, (__v8sf)__A,
+                                                    (__v8sf)__B, (__mmask8)__U);
+}
+
+static __inline__ __m256h __DEFAULT_FN_ATTRS256
+_mm256_maskz_fcmadd_pch(__mmask8 __U, __m256h __A, __m256h __B, __m256h __C) {
+  return (__m256h)__builtin_ia32_vfcmaddcph256_maskz(
+      (__v8sf)__C, (__v8sf)__A, (__v8sf)__B, (__mmask8)__U);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_fmul_pch(__m128h __A,
+                                                             __m128h __B) {
+  return (__m128h)__builtin_ia32_vfmulcph128_mask(
+      (__v4sf)__A, (__v4sf)__B, (__v4sf)_mm_undefined_ph(), (__mmask8)-1);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_mask_fmul_pch(__m128h __W,
+                                                                  __mmask8 __U,
+                                                                  __m128h __A,
+                                                                  __m128h __B) {
+  return (__m128h)__builtin_ia32_vfmulcph128_mask((__v4sf)__A, (__v4sf)__B,
+                                                  (__v4sf)__W, (__mmask8)__U);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_maskz_fmul_pch(__mmask8 __U, __m128h __A, __m128h __B) {
+  return (__m128h)__builtin_ia32_vfmulcph128_mask(
+      (__v4sf)__A, (__v4sf)__B, (__v4sf)_mm_setzero_ph(), (__mmask8)__U);
+}
+
+static __inline__ __m256h __DEFAULT_FN_ATTRS256 _mm256_fmul_pch(__m256h __A,
+                                                                __m256h __B) {
+  return (__m256h)__builtin_ia32_vfmulcph256_mask(
+      (__v8sf)__A, (__v8sf)__B, (__v8sf)_mm256_undefined_ph(), (__mmask8)-1);
+}
+
+static __inline__ __m256h __DEFAULT_FN_ATTRS256
+_mm256_mask_fmul_pch(__m256h __W, __mmask8 __U, __m256h __A, __m256h __B) {
+  return (__m256h)__builtin_ia32_vfmulcph256_mask((__v8sf)__A, (__v8sf)__B,
+                                                  (__v8sf)__W, (__mmask8)__U);
+}
+
+static __inline__ __m256h __DEFAULT_FN_ATTRS256
+_mm256_maskz_fmul_pch(__mmask8 __U, __m256h __A, __m256h __B) {
+  return (__m256h)__builtin_ia32_vfmulcph256_mask(
+      (__v8sf)__A, (__v8sf)__B, (__v8sf)_mm256_setzero_ph(), (__mmask8)__U);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_fmadd_pch(__m128h __A,
+                                                              __m128h __B,
+                                                              __m128h __C) {
+  return (__m128h)__builtin_ia32_vfmaddcph128_mask((__v4sf)__C, (__v4sf)__A,
+                                                   (__v4sf)__B, (__mmask8)-1);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_mask_fmadd_pch(__m128h __A, __mmask8 __U, __m128h __B, __m128h __C) {
+  return (__m128h)__builtin_ia32_selectps_128(
+      __U,
+      __builtin_ia32_vfmaddcph128_mask((__v4sf)__C, (__v4sf)__A, (__v4sf)__B,
+                                       (__mmask8)__U),
+      (__v4sf)__A);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_mask3_fmadd_pch(__m128h __A, __m128h __B, __m128h __C, __mmask8 __U) {
+  return (__m128h)__builtin_ia32_vfmaddcph128_mask((__v4sf)__C, (__v4sf)__A,
+                                                   (__v4sf)__B, (__mmask8)__U);
+}
+
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_maskz_fmadd_pch(__mmask8 __U, __m128h __A, __m128h __B, __m128h __C) {
+  return (__m128h)__builtin_ia32_vfmaddcph128_maskz((__v4sf)__C, (__v4sf)__A,
+                                                    (__v4sf)__B, (__mmask8)__U);
+}
+
+static __inline__ __m256h __DEFAULT_FN_ATTRS256 _mm256_fmadd_pch(__m256h __A,
+                                                                 __m256h __B,
+                                                                 __m256h __C) {
+  return (__m256h)__builtin_ia32_vfmaddcph256_mask((__v8sf)__C, (__v8sf)__A,
+                                                   (__v8sf)__B, (__mmask8)-1);
+}
+
+static __inline__ __m256h __DEFAULT_FN_ATTRS256
+_mm256_mask_fmadd_pch(__m256h __A, __mmask8 __U, __m256h __B, __m256h __C) {
+  return (__m256h)__builtin_ia32_selectps_256(
+      __U,
+      __builtin_ia32_vfmaddcph256_mask((__v8sf)__C, (__v8sf)__A, (__v8sf)__B,
+                                       (__mmask8)__U),
+      (__v8sf)__A);
+}
+
+static __inline__ __m256h __DEFAULT_FN_ATTRS256
+_mm256_mask3_fmadd_pch(__m256h __A, __m256h __B, __m256h __C, __mmask8 __U) {
+  return (__m256h)__builtin_ia32_vfmaddcph256_mask((__v8sf)__C, (__v8sf)__A,
+                                                   (__v8sf)__B, (__mmask8)__U);
+}
+
+static __inline__ __m256h __DEFAULT_FN_ATTRS256
+_mm256_maskz_fmadd_pch(__mmask8 __U, __m256h __A, __m256h __B, __m256h __C) {
+  return (__m256h)__builtin_ia32_vfmaddcph256_maskz((__v8sf)__C, (__v8sf)__A,
+                                                    (__v8sf)__B, (__mmask8)__U);
+}
+
 static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_mask_blend_ph(__mmask8 __U,
                                                                   __m128h __A,
                                                                   __m128h __W) {
   return (__m128h)__builtin_ia32_selectph_128((__mmask8)__U, (__v8hf)__W,
                                               (__v8hf)__A);
 }
 
 static __inline__ __m256h __DEFAULT_FN_ATTRS256

clang/lib/Sema/SemaChecking.cpp

bool Sema::CheckX86BuiltinRoundingOrSAE(unsigned BuiltinID, CallExpr *TheCall) {
   case X86::BI__builtin_ia32_vfmaddsubps512_mask:
   case X86::BI__builtin_ia32_vfmaddsubps512_maskz:
   case X86::BI__builtin_ia32_vfmaddsubps512_mask3:
   case X86::BI__builtin_ia32_vfmsubaddps512_mask3:
   case X86::BI__builtin_ia32_vfmaddsubph512_mask:
   case X86::BI__builtin_ia32_vfmaddsubph512_maskz:
   case X86::BI__builtin_ia32_vfmaddsubph512_mask3:
   case X86::BI__builtin_ia32_vfmsubaddph512_mask3:
+  case X86::BI__builtin_ia32_vfmaddcsh_mask:
+  case X86::BI__builtin_ia32_vfmaddcph512_mask:
+  case X86::BI__builtin_ia32_vfmaddcph512_maskz:
+  case X86::BI__builtin_ia32_vfcmaddcsh_mask:
+  case X86::BI__builtin_ia32_vfcmaddcph512_mask:
+  case X86::BI__builtin_ia32_vfcmaddcph512_maskz:
+  case X86::BI__builtin_ia32_vfmulcsh_mask:
+  case X86::BI__builtin_ia32_vfmulcph512_mask:
+  case X86::BI__builtin_ia32_vfcmulcsh_mask:
+  case X86::BI__builtin_ia32_vfcmulcph512_mask:
     ArgNum = 4;
     HasRC = true;
     break;
   }
 
   llvm::APSInt Result;
 
   // We can't check the value of a dependent argument.

clang/test/CodeGen/X86/avx512fp16-builtins.c

 }
 
 __m512h test_mm512_abs_ph(__m512h a) {
   // CHECK-LABEL: @test_mm512_abs_ph
   // CHECK: and <16 x i32>
   return _mm512_abs_ph(a);
 }
 
+__m512h test_mm512_conj_pch(__m512h __A) {
+  // CHECK-LABEL: @test_mm512_conj_pch
+  // CHECK:  %{{.*}} = bitcast <32 x half> %{{.*}} to <16 x float>
+  // CHECK:  %{{.*}} = bitcast <16 x float> %{{.*}} to <16 x i32>
+  // CHECK:  %{{.*}} = bitcast <16 x float> %{{.*}} to <16 x i32>
+  // CHECK:  %{{.*}} = xor <16 x i32> %{{.*}}, %{{.*}}
+  // CHECK:  %{{.*}} = bitcast <16 x i32> %{{.*}} to <16 x float>
+  // CHECK:  %{{.*}} = bitcast <16 x float> %{{.*}} to <32 x half>
+  return _mm512_conj_pch(__A);
+}
+
+__m512h test_mm512_mask_conj_pch(__m512h __W, __mmask32 __U, __m512h __A) {
+  // CHECK-LABEL: @test_mm512_mask_conj_pch
+  // CHECK:  %{{.*}} = trunc i32 %{{.*}} to i16
+  // CHECK:  %{{.*}} = bitcast <32 x half> %{{.*}} to <16 x float>
+  // CHECK:  %{{.*}} = bitcast <16 x float> %{{.*}} to <16 x i32>
+  // CHECK:  %{{.*}} = bitcast <16 x float> %{{.*}} to <16 x i32>
+  // CHECK:  %{{.*}} = xor <16 x i32> %{{.*}}, %{{.*}}
+  // CHECK:  %{{.*}} = bitcast <16 x i32> %{{.*}} to <16 x float>
+  // CHECK:  %{{.*}} = bitcast <16 x float> %{{.*}} to <32 x half>
+  // CHECK:  %{{.*}} = bitcast <32 x half> %{{.*}} to <16 x float>
+  // CHECK:  %{{.*}} = bitcast i16 %{{.*}} to <16 x i1>
+  // CHECK:  %{{.*}} = select <16 x i1> %{{.*}}, <16 x float> %{{.*}}, <16 x float> %{{.*}}
+  // CHECK:  %{{.*}} = bitcast <16 x float> %{{.*}} to <32 x half>
+  return _mm512_mask_conj_pch(__W, __U, __A);
+}
+
+__m512h test_mm512_maskz_conj_pch(__mmask32 __U, __m512h __A) {
+  // CHECK-LABEL: @test_mm512_maskz_conj_pch
+  // CHECK:  %{{.*}} = trunc i32 %{{.*}} to i16
+  // CHECK:  %{{.*}} = bitcast <32 x half> %{{.*}} to <16 x float>
+  // CHECK:  %{{.*}} = bitcast <16 x float> %{{.*}} to <16 x i32>
+  // CHECK:  %{{.*}} = bitcast <16 x float> %{{.*}} to <16 x i32>
+  // CHECK:  %{{.*}} = xor <16 x i32> %{{.*}}, %{{.*}}
+  // CHECK:  %{{.*}} = bitcast <16 x i32> %{{.*}} to <16 x float>
+  // CHECK:  %{{.*}} = bitcast <16 x float> %{{.*}} to <32 x half>
+  // CHECK:  %{{.*}} = bitcast i16 %{{.*}} to <16 x i1>
+  // CHECK:  %{{.*}} = select <16 x i1> %{{.*}}, <16 x float> %{{.*}}, <16 x float> %{{.*}}
+  // CHECK:  %{{.*}} = bitcast <16 x float> %{{.*}} to <32 x half>
+  return _mm512_maskz_conj_pch(__U, __A);
+}
+
 __m128h test_mm_add_round_sh(__m128h __A, __m128h __B) {
   // CHECK-LABEL: @test_mm_add_round_sh
   // CHECK: @llvm.x86.avx512fp16.mask.add.sh.round
   return _mm_add_round_sh(__A, __B, _MM_FROUND_NO_EXC | _MM_FROUND_TO_ZERO);
 }
 __m128h test_mm_mask_add_round_sh(__m128h __W, __mmask8 __U, __m128h __A, __m128h __B) {
   // CHECK-LABEL: @test_mm_mask_add_round_sh
   // CHECK: @llvm.x86.avx512fp16.mask.add.sh.round
__m128h test_mm_mask3_fnmsub_round_sh(__m128h __W, __m128h __X, __m128h __Y, __mmask8 __U) {
   // CHECK-NEXT: [[C2:%.+]] = extractelement <8 x half> [[ORIGC:%.+]], i64 0
   // CHECK-NEXT: bitcast i8 %{{.*}} to <8 x i1>
   // CHECK-NEXT: extractelement <8 x i1> %{{.*}}, i64 0
   // CHECK-NEXT: [[SEL:%.+]] = select i1 %{{.*}}, half [[FMA]], half [[C2]]
   // CHECK-NEXT: insertelement <8 x half> [[ORIGC]], half [[SEL]], i64 0
   return _mm_mask3_fnmsub_round_sh(__W, __X, __Y, __U, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
 }
 
+__m128h test_mm_fcmadd_sch(__m128h __A, __m128h __B, __m128h __C) {
+  // CHECK-LABEL: @test_mm_fcmadd_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmadd.csh
+  return _mm_fcmadd_sch(__A, __B, __C);
+}
+
+__m128h test_mm_mask_fcmadd_sch(__m128h __A, __mmask8 __U, __m128h __B, __m128h __C) {
+  // CHECK-LABEL: @test_mm_mask_fcmadd_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmadd.csh
+  // CHECK:  %{{.*}} = shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+  // CHECK:  %{{.*}} = select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}
+  return _mm_mask_fcmadd_sch(__A, __U, __B, __C);
+}
+
+__m128h test_mm_maskz_fcmadd_sch(__mmask8 __U, __m128h __A, __m128h __B, __m128h __C) {
+  // CHECK-LABEL: @test_mm_maskz_fcmadd_sch
+  // CHECK: @llvm.x86.avx512fp16.maskz.vfcmadd.csh
+  return _mm_maskz_fcmadd_sch(__U, __A, __B, __C);
+}
+
+__m128h test_mm_mask3_fcmadd_sch(__m128h __A, __m128h __B, __m128h __C, __mmask8 __U) {
+  // CHECK-LABEL: @test_mm_mask3_fcmadd_sch
+  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.csh(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}, i8 %{{.*}}, i32 4)
+  // CHECK:  %{{.*}} = extractelement <4 x float> %{{.*}}, i32 0
+  // CHECK:  %{{.*}} = insertelement <4 x float> %{{.*}}, float %{{.*}}, i32 0
+  // CHECK:  %{{.*}} = bitcast <4 x float> %{{.*}} to <8 x half>
+  return _mm_mask3_fcmadd_sch(__A, __B, __C, __U);
+}
+
+__m128h test_mm_fcmadd_round_sch(__m128h __A, __m128h __B, __m128h __C) {
+  // CHECK-LABEL: @test_mm_fcmadd_round_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmadd.csh
+  return _mm_fcmadd_round_sch(__A, __B, __C, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m128h test_mm_mask_fcmadd_round_sch(__m128h __A, __mmask8 __U, __m128h __B, __m128h __C) {
+  // CHECK-LABEL: @test_mm_mask_fcmadd_round_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmadd.csh
+  // CHECK:  %{{.*}} = shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+  // CHECK:  %{{.*}} = select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}
+  return _mm_mask_fcmadd_round_sch(__A, __U, __B, __C, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m128h test_mm_maskz_fcmadd_round_sch(__mmask8 __U, __m128h __A, __m128h __B, __m128h __C) {
+  // CHECK-LABEL: @test_mm_maskz_fcmadd_round_sch
+  // CHECK: @llvm.x86.avx512fp16.maskz.vfcmadd.csh
+  return _mm_maskz_fcmadd_round_sch(__U, __A, __B, __C, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m128h test_mm_mask3_fcmadd_round_sch(__m128h __A, __m128h __B, __m128h __C, __mmask8 __U) {
+  // CHECK-LABEL: @test_mm_mask3_fcmadd_round_sch
+  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.csh(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}, i8 %{{.*}}, i32 11)
+  // CHECK:  %{{.*}} = extractelement <4 x float> %{{.*}}, i32 0
+  // CHECK:  %{{.*}} = insertelement <4 x float> %{{.*}}, float %{{.*}}, i32 0
+  // CHECK:  %{{.*}} = bitcast <4 x float> %{{.*}} to <8 x half>
+  return _mm_mask3_fcmadd_round_sch(__A, __B, __C, __U, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m128h test_mm_fmadd_sch(__m128h __A, __m128h __B, __m128h __C) {
+  // CHECK-LABEL: @test_mm_fmadd_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmadd.csh
+  return _mm_fmadd_sch(__A, __B, __C);
+}
+
+__m128h test_mm_mask_fmadd_sch(__m128h __A, __mmask8 __U, __m128h __B, __m128h __C) {
+  // CHECK-LABEL: @test_mm_mask_fmadd_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmadd.csh
+  // CHECK:  %{{.*}} = shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+  // CHECK:  %{{.*}} = select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}
+  return _mm_mask_fmadd_sch(__A, __U, __B, __C);
+}
+
+__m128h test_mm_maskz_fmadd_sch(__mmask8 __U, __m128h __A, __m128h __B, __m128h __C) {
+  // CHECK-LABEL: @test_mm_maskz_fmadd_sch
+  // CHECK: @llvm.x86.avx512fp16.maskz.vfmadd.csh
+  return _mm_maskz_fmadd_sch(__U, __A, __B, __C);
+}
+
+__m128h test_mm_fmadd_round_sch(__m128h __A, __m128h __B, __m128h __C) {
+  // CHECK-LABEL: @test_mm_fmadd_round_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmadd.csh
+  return _mm_fmadd_round_sch(__A, __B, __C, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m128h test_mm_mask_fmadd_round_sch(__m128h __A, __mmask8 __U, __m128h __B, __m128h __C) {
+  // CHECK-LABEL: @test_mm_mask_fmadd_round_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmadd.csh
+  // CHECK:  %{{.*}} = shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+  // CHECK:  %{{.*}} = select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}
+  return _mm_mask_fmadd_round_sch(__A, __U, __B, __C, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m128h test_mm_maskz_fmadd_round_sch(__mmask8 __U, __m128h __A, __m128h __B, __m128h __C) {
+  // CHECK-LABEL: @test_mm_maskz_fmadd_round_sch
+  // CHECK: @llvm.x86.avx512fp16.maskz.vfmadd.csh
+  return _mm_maskz_fmadd_round_sch(__U, __A, __B, __C, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m128h test_mm_fcmul_sch(__m128h __A, __m128h __B) {
+  // CHECK-LABEL: @test_mm_fcmul_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.csh
+  return _mm_fcmul_sch(__A, __B);
+}
+
+__m128h test_mm_mask_fcmul_sch(__m128h __W, __mmask8 __U, __m128h __A, __m128h __B) {
+  // CHECK-LABEL: @test_mm_mask_fcmul_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.csh
+  return _mm_mask_fcmul_sch(__W, __U, __A, __B);
+}
+
+__m128h test_mm_maskz_fcmul_sch(__mmask8 __U, __m128h __A, __m128h __B) {
+  // CHECK-LABEL: @test_mm_maskz_fcmul_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.csh
+  return _mm_maskz_fcmul_sch(__U, __A, __B);
+}
+
+__m128h test_mm_fcmul_round_sch(__m128h __A, __m128h __B) {
+  // CHECK-LABEL: @test_mm_fcmul_round_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.csh
+  return _mm_fcmul_round_sch(__A, __B, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m128h test_mm_mask_fcmul_round_sch(__m128h __W, __mmask8 __U, __m128h __A, __m128h __B) {
+  // CHECK-LABEL: @test_mm_mask_fcmul_round_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.csh
+  return _mm_mask_fcmul_round_sch(__W, __U, __A, __B, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m128h test_mm_maskz_fcmul_round_sch(__mmask8 __U, __m128h __A, __m128h __B) {
+  // CHECK-LABEL: @test_mm_maskz_fcmul_round_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.csh
+  return _mm_maskz_fcmul_round_sch(__U, __A, __B, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m512h test_mm512_fcmul_pch(__m512h __A, __m512h __B) {
+  // CHECK-LABEL: @test_mm512_fcmul_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.cph.512
+  return _mm512_fcmul_pch(__A, __B);
+}
+
+__m512h test_mm512_mask_fcmul_pch(__m512h __W, __mmask16 __U, __m512h __A, __m512h __B) {
+  // CHECK-LABEL: @test_mm512_mask_fcmul_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.cph.512
+  return _mm512_mask_fcmul_pch(__W, __U, __A, __B);
+}
+
+__m512h test_mm512_maskz_fcmul_pch(__mmask16 __U, __m512h __A, __m512h __B) {
+  // CHECK-LABEL: @test_mm512_maskz_fcmul_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.cph.512
+  return _mm512_maskz_fcmul_pch(__U, __A, __B);
+}
+
+__m512h test_mm512_fcmul_round_pch(__m512h __A, __m512h __B) {
+  // CHECK-LABEL: @test_mm512_fcmul_round_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.cph.512
+  return _mm512_fcmul_round_pch(__A, __B, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m512h test_mm512_mask_fcmul_round_pch(__m512h __W, __mmask16 __U, __m512h __A, __m512h __B) {
+  // CHECK-LABEL: @test_mm512_mask_fcmul_round_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.cph.512
+  return _mm512_mask_fcmul_round_pch(__W, __U, __A, __B, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m512h test_mm512_maskz_fcmul_round_pch(__mmask16 __U, __m512h __A, __m512h __B) {
+  // CHECK-LABEL: @test_mm512_maskz_fcmul_round_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.cph.512
+  return _mm512_maskz_fcmul_round_pch(__U, __A, __B, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m512h test_mm512_fcmadd_pch(__m512h __A, __m512h __B, __m512h __C) {
+  // CHECK-LABEL: @test_mm512_fcmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmadd.cph.512
+  return _mm512_fcmadd_pch(__A, __B, __C);
+}
+
+__m512h test_mm512_mask_fcmadd_pch(__m512h __A, __mmask16 __U, __m512h __B, __m512h __C) {
+  // CHECK-LABEL: @test_mm512_mask_fcmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmadd.cph.512
+  // CHECK:  %{{.*}} = select <16 x i1> %{{.*}}, <16 x float> %{{.*}}, <16 x float> %{{.*}}
+  return _mm512_mask_fcmadd_pch(__A, __U, __B, __C);
+}
+
+__m512h test_mm512_mask3_fcmadd_pch(__m512h __A, __m512h __B, __m512h __C, __mmask16 __U) {
+  // CHECK-LABEL: @test_mm512_mask3_fcmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmadd.cph.512
+  // CHECK-NOT:  %{{.*}} = select <16 x i1> %{{.*}}, <16 x float> %{{.*}}, <16 x float> %{{.*}}
+  return _mm512_mask3_fcmadd_pch(__A, __B, __C, __U);
+}
+
+__m512h test_mm512_maskz_fcmadd_pch(__mmask16 __U, __m512h __A, __m512h __B, __m512h __C) {
+  // CHECK-LABEL: @test_mm512_maskz_fcmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.maskz.vfcmadd.cph.512
+  return _mm512_maskz_fcmadd_pch(__U, __A, __B, __C);
+}
+
+__m512h test_mm512_fcmadd_round_pch(__m512h __A, __m512h __B, __m512h __C) {
+  // CHECK-LABEL: @test_mm512_fcmadd_round_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmadd.cph.512
+  return _mm512_fcmadd_round_pch(__A, __B, __C, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m512h test_mm512_mask_fcmadd_round_pch(__m512h __A, __mmask16 __U, __m512h __B, __m512h __C) {
+  // CHECK-LABEL: @test_mm512_mask_fcmadd_round_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmadd.cph.512
+  // CHECK:  %{{.*}} = select <16 x i1> %{{.*}}, <16 x float> %{{.*}}, <16 x float> %{{.*}}
+  return _mm512_mask_fcmadd_round_pch(__A, __U, __B, __C, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m512h test_mm512_mask3_fcmadd_round_pch(__m512h __A, __m512h __B, __m512h __C, __mmask16 __U) {
+  // CHECK-LABEL: @test_mm512_mask3_fcmadd_round_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmadd.cph.512
+  // CHECK-NOT:  %{{.*}} = select <16 x i1> %{{.*}}, <16 x float> %{{.*}}, <16 x float> %{{.*}}
+  return _mm512_mask3_fcmadd_round_pch(__A, __B, __C, __U, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m512h test_mm512_maskz_fcmadd_round_pch(__mmask16 __U, __m512h __A, __m512h __B, __m512h __C) {
+  // CHECK-LABEL: @test_mm512_maskz_fcmadd_round_pch
+  // CHECK: @llvm.x86.avx512fp16.maskz.vfcmadd.cph.512
+  return _mm512_maskz_fcmadd_round_pch(__U, __A, __B, __C, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m512h test_mm512_fmul_pch(__m512h __A, __m512h __B) {
+  // CHECK-LABEL: @test_mm512_fmul_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmul.cph.512
+  return _mm512_fmul_pch(__A, __B);
+}
+
+__m512h test_mm512_mask_fmul_pch(__m512h __W, __mmask16 __U, __m512h __A, __m512h __B) {
+  // CHECK-LABEL: @test_mm512_mask_fmul_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmul.cph.512
+  return _mm512_mask_fmul_pch(__W, __U, __A, __B);
+}
+
+__m512h test_mm512_maskz_fmul_pch(__mmask16 __U, __m512h __A, __m512h __B) {
+  // CHECK-LABEL: @test_mm512_maskz_fmul_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmul.cph.512
+  return _mm512_maskz_fmul_pch(__U, __A, __B);
+}
+
+__m512h test_mm512_fmul_round_pch(__m512h __A, __m512h __B) {
+  // CHECK-LABEL: @test_mm512_fmul_round_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmul.cph.512
+  return _mm512_fmul_round_pch(__A, __B, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m512h test_mm512_mask_fmul_round_pch(__m512h __W, __mmask16 __U, __m512h __A, __m512h __B) {
+  // CHECK-LABEL: @test_mm512_mask_fmul_round_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmul.cph.512
+  return _mm512_mask_fmul_round_pch(__W, __U, __A, __B, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m512h test_mm512_maskz_fmul_round_pch(__mmask16 __U, __m512h __A, __m512h __B) {
+  // CHECK-LABEL: @test_mm512_maskz_fmul_round_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmul.cph.512
+  return _mm512_maskz_fmul_round_pch(__U, __A, __B, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m512h test_mm512_fmadd_pch(__m512h __A, __m512h __B, __m512h __C) {
+  // CHECK-LABEL: @test_mm512_fmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmadd.cph.512
+  return _mm512_fmadd_pch(__A, __B, __C);
+}
+
+__m512h test_mm512_mask_fmadd_pch(__m512h __A, __mmask16 __U, __m512h __B, __m512h __C) {
+  // CHECK-LABEL: @test_mm512_mask_fmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmadd.cph.512
+  // CHECK:  %{{.*}} = select <16 x i1> %{{.*}}, <16 x float> %{{.*}}, <16 x float> %{{.*}}
+  return _mm512_mask_fmadd_pch(__A, __U, __B, __C);
+}
+
+__m512h test_mm512_mask3_fmadd_pch(__m512h __A, __m512h __B, __m512h __C, __mmask16 __U) {
+  // CHECK-LABEL: @test_mm512_mask3_fmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmadd.cph.512
+  // CHECK-NOT:  %{{.*}} = select <16 x i1> %{{.*}}, <16 x float> %{{.*}}, <16 x float> %{{.*}}
+  return _mm512_mask3_fmadd_pch(__A, __B, __C, __U);
+}
+
+__m512h test_mm512_maskz_fmadd_pch(__mmask16 __U, __m512h __A, __m512h __B, __m512h __C) {
+  // CHECK-LABEL: @test_mm512_maskz_fmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.maskz.vfmadd.cph.512
+  return _mm512_maskz_fmadd_pch(__U, __A, __B, __C);
+}
+
+__m512h test_mm512_fmadd_round_pch(__m512h __A, __m512h __B, __m512h __C) {
+  // CHECK-LABEL: @test_mm512_fmadd_round_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmadd.cph.512
+  return _mm512_fmadd_round_pch(__A, __B, __C, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m512h test_mm512_mask_fmadd_round_pch(__m512h __A, __mmask16 __U, __m512h __B, __m512h __C) {
+  // CHECK-LABEL: @test_mm512_mask_fmadd_round_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmadd.cph.512
+  // CHECK:  %{{.*}} = select <16 x i1> %{{.*}}, <16 x float> %{{.*}}, <16 x float> %{{.*}}
+  return _mm512_mask_fmadd_round_pch(__A, __U, __B, __C, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m512h test_mm512_mask3_fmadd_round_pch(__m512h __A, __m512h __B, __m512h __C, __mmask16 __U) {
+  // CHECK-LABEL: @test_mm512_mask3_fmadd_round_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmadd.cph.512
+  // CHECK-NOT:  %{{.*}} = select <16 x i1> %{{.*}}, <16 x float> %{{.*}}, <16 x float> %{{.*}}
+  return _mm512_mask3_fmadd_round_pch(__A, __B, __C, __U, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m512h test_mm512_maskz_fmadd_round_pch(__mmask16 __U, __m512h __A, __m512h __B, __m512h __C) {
+  // CHECK-LABEL: @test_mm512_maskz_fmadd_round_pch
+  // CHECK: @llvm.x86.avx512fp16.maskz.vfmadd.cph.512
+  return _mm512_maskz_fmadd_round_pch(__U, __A, __B, __C, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m128h test_mm_fmul_sch(__m128h __A, __m128h __B) {
+  // CHECK-LABEL: @test_mm_fmul_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmul.csh
+  return _mm_fmul_sch(__A, __B);
+}
+
+__m128h test_mm_mask_fmul_sch(__m128h __W, __mmask8 __U, __m128h __A, __m128h __B) {
+  // CHECK-LABEL: @test_mm_mask_fmul_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmul.csh
+  return _mm_mask_fmul_sch(__W, __U, __A, __B);
+}
+
+__m128h test_mm_maskz_fmul_sch(__mmask8 __U, __m128h __A, __m128h __B) {
+  // CHECK-LABEL: @test_mm_maskz_fmul_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmul.csh
+  return _mm_maskz_fmul_sch(__U, __A, __B);
+}
+
+__m128h test_mm_fmul_round_sch(__m128h __A, __m128h __B) {
+  // CHECK-LABEL: @test_mm_fmul_round_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmul.csh
+  return _mm_fmul_round_sch(__A, __B, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m128h test_mm_mask_fmul_round_sch(__m128h __W, __mmask8 __U, __m128h __A, __m128h __B) {
+  // CHECK-LABEL: @test_mm_mask_fmul_round_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmul.csh
+  return _mm_mask_fmul_round_sch(__W, __U, __A, __B, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
+__m128h test_mm_maskz_fmul_round_sch(__mmask8 __U, __m128h __A, __m128h __B) {
+  // CHECK-LABEL: @test_mm_maskz_fmul_round_sch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmul.csh
+  return _mm_maskz_fmul_round_sch(__U, __A, __B, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
 _Float16 test_mm512_reduce_add_ph(__m512h __W) {
   // CHECK-LABEL: @test_mm512_reduce_add_ph
   // CHECK: call reassoc half @llvm.vector.reduce.fadd.v32f16(half 0xH8000, <32 x half> %{{.*}})
   return _mm512_reduce_add_ph(__W);
 }
 
 _Float16 test_mm512_reduce_mul_ph(__m512h __W) {
   // CHECK-LABEL: @test_mm512_reduce_mul_ph

clang/test/CodeGen/X86/avx512vlfp16-builtins.c

 }
 
 __m256h test_mm256_abs_ph(__m256h a) {
   // CHECK-LABEL: @test_mm256_abs_ph
   // CHECK: and <8 x i32>
   return _mm256_abs_ph(a);
 }
 
+__m256h test_mm256_conj_pch(__m256h __A) {
+  // CHECK-LABEL: @test_mm256_conj_pch
+  // CHECK:  %{{.*}} = bitcast <16 x half> %{{.*}} to <8 x float>
+  // CHECK:  %{{.*}} = bitcast <8 x float> %{{.*}} to <8 x i32>
+  // CHECK:  %{{.*}} = bitcast <8 x float> %{{.*}} to <8 x i32>
+  // CHECK:  %{{.*}} = xor <8 x i32> %{{.*}}, %{{.*}}
+  // CHECK:  %{{.*}} = bitcast <8 x i32> %{{.*}} to <8 x float>
+  // CHECK:  %{{.*}} = bitcast <8 x float> %{{.*}} to <16 x half>
+  return _mm256_conj_pch(__A);
+}
+
+__m256h test_mm256_mask_conj_pch(__m256h __W, __mmask32 __U, __m256h __A) {
+  // CHECK-LABEL: @test_mm256_mask_conj_pch
+  // CHECK:  %{{.*}} = trunc i32 %{{.*}} to i8
+  // CHECK:  %{{.*}} = bitcast <16 x half> %{{.*}} to <8 x float>
+  // CHECK:  %{{.*}} = bitcast <8 x float> %{{.*}} to <8 x i32>
+  // CHECK:  %{{.*}} = bitcast <8 x float> %{{.*}} to <8 x i32>
+  // CHECK:  %{{.*}} = xor <8 x i32> %{{.*}}, %{{.*}}
+  // CHECK:  %{{.*}} = bitcast <8 x i32> %{{.*}} to <8 x float>
+  // CHECK:  %{{.*}} = bitcast <8 x float> %{{.*}} to <16 x half>
+  // CHECK:  %{{.*}} = bitcast <16 x half> %{{.*}} to <8 x float>
+  // CHECK:  %{{.*}} = bitcast i8 %{{.*}} to <8 x i1>
+  // CHECK:  %{{.*}} = select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}
+  // CHECK:  %{{.*}} = bitcast <8 x float> %{{.*}} to <16 x half>
+  return _mm256_mask_conj_pch(__W, __U, __A);
+}
+
+__m256h test_mm256_maskz_conj_pch(__mmask32 __U, __m256h __A) {
+  // CHECK-LABEL: @test_mm256_maskz_conj_pch
+  // CHECK:  %{{.*}} = trunc i32 %{{.*}} to i8
+  // CHECK:  %{{.*}} = bitcast <16 x half> %{{.*}} to <8 x float>
+  // CHECK:  %{{.*}} = bitcast <8 x float> %{{.*}} to <8 x i32>
+  // CHECK:  %{{.*}} = bitcast <8 x float> %{{.*}} to <8 x i32>
+  // CHECK:  %{{.*}} = xor <8 x i32> %{{.*}}, %{{.*}}
+  // CHECK:  %{{.*}} = bitcast <8 x i32> %{{.*}} to <8 x float>
+  // CHECK:  %{{.*}} = bitcast <8 x float> %{{.*}} to <16 x half>
+  // CHECK:  %{{.*}} = bitcast i8 %{{.*}} to <8 x i1>
+  // CHECK:  %{{.*}} = select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}
+  // CHECK:  %{{.*}} = bitcast <8 x float> %{{.*}} to <16 x half>
+  return _mm256_maskz_conj_pch(__U, __A);
+}
+
+__m128h test_mm_conj_pch(__m128h __A) {
+  // CHECK-LABEL: @test_mm_conj_pch
+  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = bitcast <4 x float> %{{.*}} to <4 x i32>
+  // CHECK:  %{{.*}} = bitcast <4 x float> %{{.*}} to <4 x i32>
+  // CHECK:  %{{.*}} = xor <4 x i32> %{{.*}}, %{{.*}}
+  // CHECK:  %{{.*}} = bitcast <4 x i32> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = bitcast <4 x float> %{{.*}} to <8 x half>
+  return _mm_conj_pch(__A);
+}
+
+__m128h test_mm_mask_conj_pch(__m128h __W, __mmask32 __U, __m128h __A) {
+  // CHECK-LABEL: @test_mm_mask_conj_pch
+  // CHECK:  %{{.*}} = trunc i32 %{{.*}} to i8
+  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = bitcast <4 x float> %{{.*}} to <4 x i32>
+  // CHECK:  %{{.*}} = bitcast <4 x float> %{{.*}} to <4 x i32>
+  // CHECK:  %{{.*}} = xor <4 x i32> %{{.*}}, %{{.*}}
+  // CHECK:  %{{.*}} = bitcast <4 x i32> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = bitcast <4 x float> %{{.*}} to <8 x half>
+  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = bitcast i8 %{{.*}} to <8 x i1>
+  // CHECK:  %{{.*}} = shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+  // CHECK:  %{{.*}} = select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}
+  // CHECK:  %{{.*}} = bitcast <4 x float> %{{.*}} to <8 x half>
+  return _mm_mask_conj_pch(__W, __U, __A);
+}
+
+__m128h test_mm_maskz_conj_pch(__mmask32 __U, __m128h __A) {
+  // CHECK-LABEL: @test_mm_maskz_conj_pch
+  // CHECK:  %{{.*}} = trunc i32 %{{.*}} to i8
+  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = bitcast <4 x float> %{{.*}} to <4 x i32>
+  // CHECK:  %{{.*}} = bitcast <4 x float> %{{.*}} to <4 x i32>
+  // CHECK:  %{{.*}} = xor <4 x i32> %{{.*}}, %{{.*}}
+  // CHECK:  %{{.*}} = bitcast <4 x i32> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = bitcast <4 x float> %{{.*}} to <8 x half>
+  // CHECK:  %{{.*}} = bitcast i8 %{{.*}} to <8 x i1>
+  // CHECK:  %{{.*}} = shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+  // CHECK:  %{{.*}} = select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}
+  // CHECK:  %{{.*}} = bitcast <4 x float> %{{.*}} to <8 x half>
+  return _mm_maskz_conj_pch(__U, __A);
+}
+
 __mmask16 test_mm256_cmp_ph_mask_eq_oq(__m256h a, __m256h b) {
   // CHECK-LABEL: @test_mm256_cmp_ph_mask_eq_oq
   // CHECK: fcmp oeq <16 x half> %{{.*}}, %{{.*}}
   return _mm256_cmp_ph_mask(a, b, _CMP_EQ_OQ);
 }
 
 __mmask16 test_mm256_cmp_ph_mask_lt_os(__m256h a, __m256h b) {
   // CHECK-LABEL: test_mm256_cmp_ph_mask_lt_os

 
 __m256h test_mm256_mask3_fnmsub_ph(__m256h __A, __m256h __B, __m256h __C, __mmask16 __U) {
   // CHECK-LABEL: @test_mm256_mask3_fnmsub_ph
   // CHECK: fneg
   // CHECK: fneg
   // CHECK: call <16 x half> @llvm.fma.v16f16(<16 x half> %{{.*}}, <16 x half> %{{.*}}, <16 x half> %{{.*}})
   return _mm256_mask3_fnmsub_ph(__A, __B, __C, __U);
 }
+
+__m128h test_mm_fcmul_pch(__m128h __A, __m128h __B) {
+  // CHECK-LABEL: @test_mm_fcmul_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.cph.128
+  return _mm_fcmul_pch(__A, __B);
+}
+
+__m128h test_mm_mask_fcmul_pch(__m128h __W, __mmask8 __U, __m128h __A, __m128h __B) {
+  // CHECK-LABEL: @test_mm_mask_fcmul_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.cph.128
+  return _mm_mask_fcmul_pch(__W, __U, __A, __B);
+}
+
+__m128h test_mm_maskz_fcmul_pch(__mmask8 __U, __m128h __A, __m128h __B) {
+  // CHECK-LABEL: @test_mm_maskz_fcmul_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.cph.128
+  return _mm_maskz_fcmul_pch(__U, __A, __B);
+}
+
+__m256h test_mm256_fcmul_pch(__m256h __A, __m256h __B) {
+  // CHECK-LABEL: @test_mm256_fcmul_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.cph.256
+  return _mm256_fcmul_pch(__A, __B);
+}
+
+__m256h test_mm256_mask_fcmul_pch(__m256h __W, __mmask8 __U, __m256h __A, __m256h __B) {
+  // CHECK-LABEL: @test_mm256_mask_fcmul_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.cph.256
+  return _mm256_mask_fcmul_pch(__W, __U, __A, __B);
+}
+
+__m256h test_mm256_maskz_fcmul_pch(__mmask8 __U, __m256h __A, __m256h __B) {
+  // CHECK-LABEL: @test_mm256_maskz_fcmul_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.cph.256
+  return _mm256_maskz_fcmul_pch(__U, __A, __B);
+}
+
+__m128h test_mm_fcmadd_pch(__m128h __A, __m128h __B, __m128h __C) {
+  // CHECK-LABEL: @test_mm_fcmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmadd.cph.128
+  return _mm_fcmadd_pch(__A, __B, __C);
+}
+
+__m128h test_mm_mask_fcmadd_pch(__m128h __A, __mmask8 __U, __m128h __B, __m128h __C) {
+  // CHECK-LABEL: @test_mm_mask_fcmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmadd.cph.128
+  // CHECK:  %{{.*}} = shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+  // CHECK:  %{{.*}} = select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}
+  return _mm_mask_fcmadd_pch(__A, __U, __B, __C);
+}
+
+__m128h test_mm_mask3_fcmadd_pch(__m128h __A, __m128h __B, __m128h __C, __mmask8 __U) {
+  // CHECK-LABEL: @test_mm_mask3_fcmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmadd.cph.128
+  // CHECK-NOT:  %{{.*}} = select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}
+  return _mm_mask3_fcmadd_pch(__A, __B, __C, __U);
+}
+
+__m128h test_mm_maskz_fcmadd_pch(__mmask8 __U, __m128h __A, __m128h __B, __m128h __C) {
+  // CHECK-LABEL: @test_mm_maskz_fcmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.maskz.vfcmadd.cph.128
+  return _mm_maskz_fcmadd_pch(__U, __A, __B, __C);
+}
+
+__m256h test_mm256_fcmadd_pch(__m256h __A, __m256h __B, __m256h __C) {
+  // CHECK-LABEL: @test_mm256_fcmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmadd.cph.256
+  return _mm256_fcmadd_pch(__A, __B, __C);
+}
+
+__m256h test_mm256_mask_fcmadd_pch(__m256h __A, __mmask8 __U, __m256h __B, __m256h __C) {
+  // CHECK-LABEL: @test_mm256_mask_fcmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmadd.cph.256
+  // CHECK:  %{{.*}} = select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}
+  return _mm256_mask_fcmadd_pch(__A, __U, __B, __C);
+}
+
+__m256h test_mm256_mask3_fcmadd_pch(__m256h __A, __m256h __B, __m256h __C, __mmask8 __U) {
+  // CHECK-LABEL: @test_mm256_mask3_fcmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfcmadd.cph.256
+  // CHECK-NOT:  %{{.*}} = select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}
+  return _mm256_mask3_fcmadd_pch(__A, __B, __C, __U);
+}
+
+__m256h test_mm256_maskz_fcmadd_pch(__mmask8 __U, __m256h __A, __m256h __B, __m256h __C) {
+  // CHECK-LABEL: @test_mm256_maskz_fcmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.maskz.vfcmadd.cph.256
+  return _mm256_maskz_fcmadd_pch(__U, __A, __B, __C);
+}
+
+__m128h test_mm_fmul_pch(__m128h __A, __m128h __B) {
+  // CHECK-LABEL: @test_mm_fmul_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmul.cph.128
+  return _mm_fmul_pch(__A, __B);
+}
+
+__m128h test_mm_mask_fmul_pch(__m128h __W, __mmask8 __U, __m128h __A, __m128h __B) {
+  // CHECK-LABEL: @test_mm_mask_fmul_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmul.cph.128
+  return _mm_mask_fmul_pch(__W, __U, __A, __B);
+}
+
+__m128h test_mm_maskz_fmul_pch(__mmask8 __U, __m128h __A, __m128h __B) {
+  // CHECK-LABEL: @test_mm_maskz_fmul_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmul.cph.128
+  return _mm_maskz_fmul_pch(__U, __A, __B);
+}
+
+__m256h test_mm256_fmul_pch(__m256h __A, __m256h __B) {
+  // CHECK-LABEL: @test_mm256_fmul_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmul.cph.256
+  return _mm256_fmul_pch(__A, __B);
+}
+
+__m256h test_mm256_mask_fmul_pch(__m256h __W, __mmask8 __U, __m256h __A, __m256h __B) {
+  // CHECK-LABEL: @test_mm256_mask_fmul_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmul.cph.256
+  return _mm256_mask_fmul_pch(__W, __U, __A, __B);
+}
+
+__m256h test_mm256_maskz_fmul_pch(__mmask8 __U, __m256h __A, __m256h __B) {
+  // CHECK-LABEL: @test_mm256_maskz_fmul_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmul.cph.256
+  return _mm256_maskz_fmul_pch(__U, __A, __B);
+}
+
+__m128h test_mm_fmadd_pch(__m128h __A, __m128h __B, __m128h __C) {
+  // CHECK-LABEL: @test_mm_fmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmadd.cph.128
+  return _mm_fmadd_pch(__A, __B, __C);
+}
+
+__m128h test_mm_mask_fmadd_pch(__m128h __A, __mmask8 __U, __m128h __B, __m128h __C) {
+  // CHECK-LABEL: @test_mm_mask_fmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmadd.cph.128
+  // CHECK:  %{{.*}} = shufflevector <8 x i1> %{{.*}}, <8 x i1> %{{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+  // CHECK:  %{{.*}} = select <4 x i1> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}
+  return _mm_mask_fmadd_pch(__A, __U, __B, __C);
+}
+
+__m128h test_mm_mask3_fmadd_pch(__m128h __A, __m128h __B, __m128h __C, __mmask8 __U) {
+  // CHECK-LABEL: @test_mm_mask3_fmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmadd.cph.128
+  return _mm_mask3_fmadd_pch(__A, __B, __C, __U);
+}
+
+__m128h test_mm_maskz_fmadd_pch(__mmask8 __U, __m128h __A, __m128h __B, __m128h __C) {
+  // CHECK-LABEL: @test_mm_maskz_fmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.maskz.vfmadd.cph.128
+  return _mm_maskz_fmadd_pch(__U, __A, __B, __C);
+}
+
+__m256h test_mm256_fmadd_pch(__m256h __A, __m256h __B, __m256h __C) {
+  // CHECK-LABEL: @test_mm256_fmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmadd.cph.256
+  return _mm256_fmadd_pch(__A, __B, __C);
+}
+
+__m256h test_mm256_mask_fmadd_pch(__m256h __A, __mmask8 __U, __m256h __B, __m256h __C) {
+  // CHECK-LABEL: @test_mm256_mask_fmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmadd.cph.256
+  // CHECK:  %{{.*}} = select <8 x i1> %{{.*}}, <8 x float> %{{.*}}, <8 x float> %{{.*}}
+  return _mm256_mask_fmadd_pch(__A, __U, __B, __C);
+}
+
+__m256h test_mm256_mask3_fmadd_pch(__m256h __A, __m256h __B, __m256h __C, __mmask8 __U) {
+  // CHECK-LABEL: @test_mm256_mask3_fmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.mask.vfmadd.cph.256
+  return _mm256_mask3_fmadd_pch(__A, __B, __C, __U);
+}
+
+__m256h test_mm256_maskz_fmadd_pch(__mmask8 __U, __m256h __A, __m256h __B, __m256h __C) {
+  // CHECK-LABEL: @test_mm256_maskz_fmadd_pch
+  // CHECK: @llvm.x86.avx512fp16.maskz.vfmadd.cph.256
+  return _mm256_maskz_fmadd_pch(__U, __A, __B, __C);
+}
+
 __m128h test_mm_mask_blend_ph(__mmask8 __U, __m128h __A, __m128h __W) {
   // CHECK-LABEL: @test_mm_mask_blend_ph
   // CHECK:  %{{.*}} = bitcast i8 %{{.*}} to <8 x i1>
   // CHECK:  %{{.*}} = select <8 x i1> %{{.*}}, <8 x half> %{{.*}}, <8 x half> %{{.*}}
   return _mm_mask_blend_ph(__U, __A, __W);
 }
 
 __m256h test_mm256_mask_blend_ph(__mmask16 __U, __m256h __A, __m256h __W) {

llvm/include/llvm/IR/IntrinsicsX86.td

let TargetPrefix = "x86" in {
   def int_x86_avx512fp16_vfmaddsub_ph_512
       : Intrinsic<[ llvm_v32f16_ty ],
                   [ llvm_v32f16_ty, llvm_v32f16_ty, llvm_v32f16_ty, llvm_i32_ty ],
                   [ IntrNoMem, ImmArg<ArgIndex<3>> ]>;
   def int_x86_avx512fp16_vfmadd_f16
       : Intrinsic<[ llvm_half_ty ],
                   [ llvm_half_ty, llvm_half_ty, llvm_half_ty, llvm_i32_ty ],
                   [ IntrNoMem, ImmArg<ArgIndex<3>> ]>;
+
+  def int_x86_avx512fp16_mask_vfcmadd_cph_128
+      : GCCBuiltin<"__builtin_ia32_vfcmaddcph128_mask">,
+        Intrinsic<[ llvm_v4f32_ty ],
+                  [ llvm_v4f32_ty, llvm_v4f32_ty, llvm_v4f32_ty, llvm_i8_ty ],
+                  [ IntrNoMem ]>;
+  def int_x86_avx512fp16_maskz_vfcmadd_cph_128
+      : GCCBuiltin<"__builtin_ia32_vfcmaddcph128_maskz">,
+        Intrinsic<[ llvm_v4f32_ty ],
+                  [ llvm_v4f32_ty, llvm_v4f32_ty, llvm_v4f32_ty, llvm_i8_ty ],
+                  [ IntrNoMem ]>;
+  def int_x86_avx512fp16_mask_vfcmadd_cph_256
+      : GCCBuiltin<"__builtin_ia32_vfcmaddcph256_mask">,
+        Intrinsic<[ llvm_v8f32_ty ],
+                  [ llvm_v8f32_ty, llvm_v8f32_ty, llvm_v8f32_ty, llvm_i8_ty ],
+                  [ IntrNoMem ]>;
+  def int_x86_avx512fp16_maskz_vfcmadd_cph_256
+      : GCCBuiltin<"__builtin_ia32_vfcmaddcph256_maskz">,
+        Intrinsic<[ llvm_v8f32_ty ],
+                  [ llvm_v8f32_ty, llvm_v8f32_ty, llvm_v8f32_ty, llvm_i8_ty ],
+                  [ IntrNoMem ]>;
+  def int_x86_avx512fp16_mask_vfcmadd_cph_512
+      : GCCBuiltin<"__builtin_ia32_vfcmaddcph512_mask">,
+        Intrinsic<[ llvm_v16f32_ty ],
+                  [ llvm_v16f32_ty, llvm_v16f32_ty, llvm_v16f32_ty, llvm_i16_ty,
+                    llvm_i32_ty ],
+                  [ IntrNoMem, ImmArg<ArgIndex<4>> ]>;
+  def int_x86_avx512fp16_maskz_vfcmadd_cph_512
+      : GCCBuiltin<"__builtin_ia32_vfcmaddcph512_maskz">,
+        Intrinsic<[ llvm_v16f32_ty ],
+                  [ llvm_v16f32_ty, llvm_v16f32_ty, llvm_v16f32_ty, llvm_i16_ty,
+                    llvm_i32_ty ],
+                  [ IntrNoMem, ImmArg<ArgIndex<4>> ]>;
+  def int_x86_avx512fp16_mask_vfmadd_cph_128
+      : GCCBuiltin<"__builtin_ia32_vfmaddcph128_mask">,
+        Intrinsic<[ llvm_v4f32_ty ],
+                  [ llvm_v4f32_ty, llvm_v4f32_ty, llvm_v4f32_ty, llvm_i8_ty ],
+                  [ IntrNoMem ]>;
+  def int_x86_avx512fp16_maskz_vfmadd_cph_128
+      : GCCBuiltin<"__builtin_ia32_vfmaddcph128_maskz">,
+        Intrinsic<[ llvm_v4f32_ty ],
+                  [ llvm_v4f32_ty, llvm_v4f32_ty, llvm_v4f32_ty, llvm_i8_ty ],
+                  [ IntrNoMem ]>;
+  def int_x86_avx512fp16_mask_vfmadd_cph_256
+      : GCCBuiltin<"__builtin_ia32_vfmaddcph256_mask">,
+        Intrinsic<[ llvm_v8f32_ty ],
+                  [ llvm_v8f32_ty, llvm_v8f32_ty, llvm_v8f32_ty, llvm_i8_ty ],
+                  [ IntrNoMem ]>;
+  def int_x86_avx512fp16_maskz_vfmadd_cph_256
+      : GCCBuiltin<"__builtin_ia32_vfmaddcph256_maskz">,
+        Intrinsic<[ llvm_v8f32_ty ],
+                  [ llvm_v8f32_ty, llvm_v8f32_ty, llvm_v8f32_ty, llvm_i8_ty ],
+                  [ IntrNoMem ]>;
+  def int_x86_avx512fp16_mask_vfmadd_cph_512
+      : GCCBuiltin<"__builtin_ia32_vfmaddcph512_mask">,
+        Intrinsic<[ llvm_v16f32_ty ],
+                  [ llvm_v16f32_ty, llvm_v16f32_ty, llvm_v16f32_ty, llvm_i16_ty,
+                    llvm_i32_ty ],
+                  [ IntrNoMem, ImmArg<ArgIndex<4>> ]>;
+  def int_x86_avx512fp16_maskz_vfmadd_cph_512
+      : GCCBuiltin<"__builtin_ia32_vfmaddcph512_maskz">,
+        Intrinsic<[ llvm_v16f32_ty ],
+                  [ llvm_v16f32_ty, llvm_v16f32_ty, llvm_v16f32_ty, llvm_i16_ty,
+                    llvm_i32_ty ],
+                  [ IntrNoMem, ImmArg<ArgIndex<4>> ]>;
+  def int_x86_avx512fp16_mask_vfmadd_csh
+      : GCCBuiltin<"__builtin_ia32_vfmaddcsh_mask">,
+        Intrinsic<[ llvm_v4f32_ty ],
+                  [ llvm_v4f32_ty, llvm_v4f32_ty, llvm_v4f32_ty, llvm_i8_ty,
+                    llvm_i32_ty ],
+                  [ IntrNoMem, ImmArg<ArgIndex<4>> ]>;
+  def int_x86_avx512fp16_maskz_vfmadd_csh
+      : GCCBuiltin<"__builtin_ia32_vfmaddcsh_maskz">,
+        Intrinsic<[ llvm_v4f32_ty ],
+                  [ llvm_v4f32_ty, llvm_v4f32_ty, llvm_v4f32_ty, llvm_i8_ty,
+                    llvm_i32_ty ],
+                  [ IntrNoMem, ImmArg<ArgIndex<4>> ]>;
+  def int_x86_avx512fp16_mask_vfcmadd_csh
+      : GCCBuiltin<"__builtin_ia32_vfcmaddcsh_mask">,
+        Intrinsic<[ llvm_v4f32_ty ],
+                  [ llvm_v4f32_ty, llvm_v4f32_ty, llvm_v4f32_ty, llvm_i8_ty,
+                    llvm_i32_ty ],
+                  [ IntrNoMem, ImmArg<ArgIndex<4>> ]>;
+  def int_x86_avx512fp16_maskz_vfcmadd_csh
+      : GCCBuiltin<"__builtin_ia32_vfcmaddcsh_maskz">,
+        Intrinsic<[ llvm_v4f32_ty ],
+                  [ llvm_v4f32_ty, llvm_v4f32_ty, llvm_v4f32_ty, llvm_i8_ty,
+                    llvm_i32_ty ],
+                  [ IntrNoMem, ImmArg<ArgIndex<4>> ]>;
+  def int_x86_avx512fp16_mask_vfmul_cph_128
+      : GCCBuiltin<"__builtin_ia32_vfmulcph128_mask">,
+        Intrinsic<[ llvm_v4f32_ty ],
+                  [ llvm_v4f32_ty, llvm_v4f32_ty, llvm_v4f32_ty, llvm_i8_ty ],
+                  [ IntrNoMem ]>;
+  def int_x86_avx512fp16_mask_vfcmul_cph_128
+      : GCCBuiltin<"__builtin_ia32_vfcmulcph128_mask">,
+        Intrinsic<[ llvm_v4f32_ty ],
+                  [ llvm_v4f32_ty, llvm_v4f32_ty, llvm_v4f32_ty, llvm_i8_ty ],
+                  [ IntrNoMem ]>;
+  def int_x86_avx512fp16_mask_vfmul_cph_256
+      : GCCBuiltin<"__builtin_ia32_vfmulcph256_mask">,
+        Intrinsic<[ llvm_v8f32_ty ],
+                  [ llvm_v8f32_ty, llvm_v8f32_ty, llvm_v8f32_ty, llvm_i8_ty ],
+                  [ IntrNoMem ]>;
+  def int_x86_avx512fp16_mask_vfcmul_cph_256
+      : GCCBuiltin<"__builtin_ia32_vfcmulcph256_mask">,
+        Intrinsic<[ llvm_v8f32_ty ],
+                  [ llvm_v8f32_ty, llvm_v8f32_ty, llvm_v8f32_ty, llvm_i8_ty ],
+                  [ IntrNoMem ]>;
+  def int_x86_avx512fp16_mask_vfmul_cph_512
+      : GCCBuiltin<"__builtin_ia32_vfmulcph512_mask">,
+        Intrinsic<[ llvm_v16f32_ty ],
+                  [ llvm_v16f32_ty, llvm_v16f32_ty, llvm_v16f32_ty, llvm_i16_ty,
+                    llvm_i32_ty ],
+                  [ IntrNoMem, ImmArg<ArgIndex<4>> ]>;
+  def int_x86_avx512fp16_mask_vfcmul_cph_512
+      : GCCBuiltin<"__builtin_ia32_vfcmulcph512_mask">,
+        Intrinsic<[ llvm_v16f32_ty ],
+                  [ llvm_v16f32_ty, llvm_v16f32_ty, llvm_v16f32_ty, llvm_i16_ty,
+                    llvm_i32_ty ],
+                  [ IntrNoMem, ImmArg<ArgIndex<4>> ]>;
+  def int_x86_avx512fp16_mask_vfmul_csh
+      : GCCBuiltin<"__builtin_ia32_vfmulcsh_mask">,
+        Intrinsic<[ llvm_v4f32_ty ],
+                  [ llvm_v4f32_ty, llvm_v4f32_ty, llvm_v4f32_ty, llvm_i8_ty,
+                    llvm_i32_ty ],
+                  [ IntrNoMem, ImmArg<ArgIndex<4>> ]>;
+  def int_x86_avx512fp16_mask_vfcmul_csh
+      : GCCBuiltin<"__builtin_ia32_vfcmulcsh_mask">,
+        Intrinsic<[ llvm_v4f32_ty ],
+                  [ llvm_v4f32_ty, llvm_v4f32_ty, llvm_v4f32_ty, llvm_i8_ty,
+                    llvm_i32_ty ],
+                  [ IntrNoMem, ImmArg<ArgIndex<4>> ]>;
 }

llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp

if (Src2Enc % 4 != 0) {
       return Warning(Ops[0]->getStartLoc(),
                      "source register '" + RegName + "' implicitly denotes '" +
                      RegName.take_front(3) + Twine(GroupStart) + "' to '" +
                      RegName.take_front(3) + Twine(GroupEnd) +
                      "' source group");
     }
     break;
   }
+  case X86::VFCMADDCPHZ128m:
+  case X86::VFCMADDCPHZ256m:
+  case X86::VFCMADDCPHZm:
+  case X86::VFCMADDCPHZ128mb:
+  case X86::VFCMADDCPHZ256mb:
+  case X86::VFCMADDCPHZmb:
+  case X86::VFCMADDCPHZ128mbk:
+  case X86::VFCMADDCPHZ256mbk:
+  case X86::VFCMADDCPHZmbk:
+  case X86::VFCMADDCPHZ128mbkz:
+  case X86::VFCMADDCPHZ256mbkz:
+  case X86::VFCMADDCPHZmbkz:
+  case X86::VFCMADDCPHZ128mk:
+  case X86::VFCMADDCPHZ256mk:
+  case X86::VFCMADDCPHZmk:
+  case X86::VFCMADDCPHZ128mkz:
+  case X86::VFCMADDCPHZ256mkz:
+  case X86::VFCMADDCPHZmkz:
+  case X86::VFCMADDCPHZ128r:
+  case X86::VFCMADDCPHZ256r:
+  case X86::VFCMADDCPHZr:
+  case X86::VFCMADDCPHZ128rk:
+  case X86::VFCMADDCPHZ256rk:
+  case X86::VFCMADDCPHZrk:
+  case X86::VFCMADDCPHZ128rkz:
+  case X86::VFCMADDCPHZ256rkz:
+  case X86::VFCMADDCPHZrkz:
+  case X86::VFCMADDCPHZrb:
+  case X86::VFCMADDCPHZrbk:
+  case X86::VFCMADDCPHZrbkz:
+  case X86::VFCMADDCSHZm:
+  case X86::VFCMADDCSHZmk:
+  case X86::VFCMADDCSHZmkz:
+  case X86::VFCMADDCSHZr:
+  case X86::VFCMADDCSHZrb:
+  case X86::VFCMADDCSHZrbk:
+  case X86::VFCMADDCSHZrbkz:
+  case X86::VFCMADDCSHZrk:
+  case X86::VFCMADDCSHZrkz:
+  case X86::VFMADDCPHZ128m:
+  case X86::VFMADDCPHZ256m:
+  case X86::VFMADDCPHZm:
+  case X86::VFMADDCPHZ128mb:
+  case X86::VFMADDCPHZ256mb:
+  case X86::VFMADDCPHZmb:
+  case X86::VFMADDCPHZ128mbk:
+  case X86::VFMADDCPHZ256mbk:
+  case X86::VFMADDCPHZmbk:
+  case X86::VFMADDCPHZ128mbkz:
+  case X86::VFMADDCPHZ256mbkz:
+  case X86::VFMADDCPHZmbkz:
+  case X86::VFMADDCPHZ128mk:
+  case X86::VFMADDCPHZ256mk:
+  case X86::VFMADDCPHZmk:
+  case X86::VFMADDCPHZ128mkz:
+  case X86::VFMADDCPHZ256mkz:
+  case X86::VFMADDCPHZmkz:
+  case X86::VFMADDCPHZ128r:
+  case X86::VFMADDCPHZ256r:
+  case X86::VFMADDCPHZr:
+  case X86::VFMADDCPHZ128rk:
+  case X86::VFMADDCPHZ256rk:
+  case X86::VFMADDCPHZrk:
+  case X86::VFMADDCPHZ128rkz:
+  case X86::VFMADDCPHZ256rkz:
+  case X86::VFMADDCPHZrkz:
+  case X86::VFMADDCPHZrb:
+  case X86::VFMADDCPHZrbk:
+  case X86::VFMADDCPHZrbkz:
+  case X86::VFMADDCSHZm:
+  case X86::VFMADDCSHZmk:
+  case X86::VFMADDCSHZmkz:
+  case X86::VFMADDCSHZr:
+  case X86::VFMADDCSHZrb:
+  case X86::VFMADDCSHZrbk:
+  case X86::VFMADDCSHZrbkz:
+  case X86::VFMADDCSHZrk:
+  case X86::VFMADDCSHZrkz: {
+    unsigned Dest = Inst.getOperand(0).getReg();
+    for (unsigned i = 2; i < Inst.getNumOperands(); i++)
+      if (Inst.getOperand(i).isReg() && Dest == Inst.getOperand(i).getReg())
+        return Warning(Ops[0]->getStartLoc(), "Destination register should be "
+                                              "distinct from source registers");
+    break;
+  }
+  case X86::VFCMULCPHZ128rm:
+  case X86::VFCMULCPHZ256rm:
+  case X86::VFCMULCPHZrm:
+  case X86::VFCMULCPHZ128rmb:
+  case X86::VFCMULCPHZ256rmb:
+  case X86::VFCMULCPHZrmb:
+  case X86::VFCMULCPHZ128rmbk:
+  case X86::VFCMULCPHZ256rmbk:
+  case X86::VFCMULCPHZrmbk:
+  case X86::VFCMULCPHZ128rmbkz:
+  case X86::VFCMULCPHZ256rmbkz:
+  case X86::VFCMULCPHZrmbkz:
+  case X86::VFCMULCPHZ128rmk:
+  case X86::VFCMULCPHZ256rmk:
+  case X86::VFCMULCPHZrmk:
+  case X86::VFCMULCPHZ128rmkz:
+  case X86::VFCMULCPHZ256rmkz:
+  case X86::VFCMULCPHZrmkz:
+  case X86::VFCMULCPHZ128rr:
+  case X86::VFCMULCPHZ256rr:
+  case X86::VFCMULCPHZrr:
+  case X86::VFCMULCPHZ128rrk:
+  case X86::VFCMULCPHZ256rrk:
+  case X86::VFCMULCPHZrrk:
+  case X86::VFCMULCPHZ128rrkz:
+  case X86::VFCMULCPHZ256rrkz:
+  case X86::VFCMULCPHZrrkz:
+  case X86::VFCMULCPHZrrb:
+  case X86::VFCMULCPHZrrbk:
+  case X86::VFCMULCPHZrrbkz:
+  case X86::VFCMULCSHZrm:
+  case X86::VFCMULCSHZrmk:
+  case X86::VFCMULCSHZrmkz:
+  case X86::VFCMULCSHZrr:
+  case X86::VFCMULCSHZrrb:
+  case X86::VFCMULCSHZrrbk:
+  case X86::VFCMULCSHZrrbkz:
+  case X86::VFCMULCSHZrrk:
+  case X86::VFCMULCSHZrrkz:
+  case X86::VFMULCPHZ128rm:
+  case X86::VFMULCPHZ256rm:
+  case X86::VFMULCPHZrm:
+  case X86::VFMULCPHZ128rmb:
+  case X86::VFMULCPHZ256rmb:
+  case X86::VFMULCPHZrmb:
+  case X86::VFMULCPHZ128rmbk:
+  case X86::VFMULCPHZ256rmbk:
+  case X86::VFMULCPHZrmbk:
+  case X86::VFMULCPHZ128rmbkz:
+  case X86::VFMULCPHZ256rmbkz:
+  case X86::VFMULCPHZrmbkz:
+  case X86::VFMULCPHZ128rmk:
+  case X86::VFMULCPHZ256rmk:
+  case X86::VFMULCPHZrmk:
+  case X86::VFMULCPHZ128rmkz:
+  case X86::VFMULCPHZ256rmkz:
+  case X86::VFMULCPHZrmkz:
+  case X86::VFMULCPHZ128rr:
+  case X86::VFMULCPHZ256rr:
+  case X86::VFMULCPHZrr:
+  case X86::VFMULCPHZ128rrk:
+  case X86::VFMULCPHZ256rrk:
+  case X86::VFMULCPHZrrk:
+  case X86::VFMULCPHZ128rrkz:
+  case X86::VFMULCPHZ256rrkz:
+  case X86::VFMULCPHZrrkz:
+  case X86::VFMULCPHZrrb:
+  case X86::VFMULCPHZrrbk:
+  case X86::VFMULCPHZrrbkz:
+  case X86::VFMULCSHZrm:
+  case X86::VFMULCSHZrmk:
+  case X86::VFMULCSHZrmkz:
+  case X86::VFMULCSHZrr:
+  case X86::VFMULCSHZrrb:
+  case X86::VFMULCSHZrrbk:
+  case X86::VFMULCSHZrrbkz:
+  case X86::VFMULCSHZrrk:
+  case X86::VFMULCSHZrrkz: {
+    unsigned Dest = Inst.getOperand(0).getReg();
+    for (unsigned i = 1; i < Inst.getNumOperands(); i++)
+      if (Inst.getOperand(i).isReg() && Dest == Inst.getOperand(i).getReg())
+        return Warning(Ops[0]->getStartLoc(), "Destination register should be "
+                                              "distinct from source registers");
+    break;
+  }
   }
 
   const MCInstrDesc &MCID = MII.get(Inst.getOpcode());
   // Check that we aren't mixing AH/BH/CH/DH with REX prefix. We only need to
   // check this with the legacy encoding, VEX/EVEX/XOP don't use REX.
   if ((MCID.TSFlags & X86II::EncodingMask) == 0) {
     MCPhysReg HReg = X86::NoRegister;
     bool UsesRex = MCID.TSFlags & X86II::REX_W;

llvm/lib/Target/X86/X86ISelLowering.h

enum NodeType : unsigned {
     // FMA with rounding mode.
     FMADD_RND,
     FNMADD_RND,
     FMSUB_RND,
     FNMSUB_RND,
     FMADDSUB_RND,
     FMSUBADD_RND,
 
+    // AVX512-FP16 complex addition and multiplication.
+    VFMADDC,
+    VFMADDC_RND,
+    VFCMADDC,
+    VFCMADDC_RND,
+
+    VFMULC,
+    VFMULC_RND,
+    VFCMULC,
+    VFCMULC_RND,
+
+    VFMADDCSH,
+    VFMADDCSH_RND,
+    VFCMADDCSH,
+    VFCMADDCSH_RND,
+
+    VFMULCSH,
+    VFMULCSH_RND,
+    VFCMULCSH,
+    VFCMULCSH_RND,
+
     // Compress and expand.
     COMPRESS,
     EXPAND,
 
     // Bits shuffle
     VPSHUFBITQMB,
 
     // Convert Unsigned/Integer to Floating-Point Value with rounding mode.

llvm/lib/Target/X86/X86ISelLowering.cpp

if (IntrData) {
     }
     case VPERM_2OP : {
       SDValue Src1 = Op.getOperand(1);
       SDValue Src2 = Op.getOperand(2);
 
       // Swap Src1 and Src2 in the node creation
       return DAG.getNode(IntrData->Opc0, dl, VT,Src2, Src1);
     }
+    case FMA_OP_MASKZ:
+    case FMA_OP_MASK: {
+      SDValue Src1 = Op.getOperand(1);
+      SDValue Src2 = Op.getOperand(2);
+      SDValue Src3 = Op.getOperand(3);
+      SDValue Mask = Op.getOperand(4);
+      MVT VT = Op.getSimpleValueType();
+
+      SDValue PassThru = Src1;
+      if (IntrData->Type == FMA_OP_MASKZ)
+        PassThru = getZeroVector(VT, Subtarget, DAG, dl);
+
+      // We add rounding mode to the Node when
+      //   - RC Opcode is specified and
+      //   - RC is not "current direction".
+      SDValue NewOp;
+      if (IntrData->Opc1 != 0) {
+        SDValue Rnd = Op.getOperand(5);
+        unsigned RC = 0;
+        if (isRoundModeSAEToX(Rnd, RC))
+          NewOp = DAG.getNode(IntrData->Opc1, dl, VT, Src1, Src2, Src3,
+                              DAG.getTargetConstant(RC, dl, MVT::i32));
+        else if (!isRoundModeCurDirection(Rnd))
+          return SDValue();
+      }
+      if (!NewOp)
+        NewOp = DAG.getNode(IntrData->Opc0, dl, VT, Src1, Src2, Src3);
+      return getVectorMaskingNode(NewOp, Mask, PassThru, Subtarget, DAG);
+    }
     case IFMA_OP:
       // NOTE: We need to swizzle the operands to pass the multiply operands
       // first.
       return DAG.getNode(IntrData->Opc0, dl, Op.getValueType(),
                          Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
     case FPCLASSS: {
       SDValue Src1 = Op.getOperand(1);
       SDValue Imm = Op.getOperand(2);
#define NODE_NAME_CASE(NODE) case X86ISD::NODE: return "X86ISD::" #NODE;
   NODE_NAME_CASE(FMADDSUB)
   NODE_NAME_CASE(FMSUBADD)
   NODE_NAME_CASE(FMADD_RND)
   NODE_NAME_CASE(FNMADD_RND)
   NODE_NAME_CASE(FMSUB_RND)
   NODE_NAME_CASE(FNMSUB_RND)
   NODE_NAME_CASE(FMADDSUB_RND)
   NODE_NAME_CASE(FMSUBADD_RND)
+  NODE_NAME_CASE(VFMADDC)
+  NODE_NAME_CASE(VFMADDC_RND)
+  NODE_NAME_CASE(VFCMADDC)
+  NODE_NAME_CASE(VFCMADDC_RND)
+  NODE_NAME_CASE(VFMULC)
+  NODE_NAME_CASE(VFMULC_RND)
+  NODE_NAME_CASE(VFCMULC)
+  NODE_NAME_CASE(VFCMULC_RND)
+  NODE_NAME_CASE(VFMULCSH)
+  NODE_NAME_CASE(VFMULCSH_RND)
+  NODE_NAME_CASE(VFCMULCSH)
+  NODE_NAME_CASE(VFCMULCSH_RND)
+  NODE_NAME_CASE(VFMADDCSH)
+  NODE_NAME_CASE(VFMADDCSH_RND)
+  NODE_NAME_CASE(VFCMADDCSH)
+  NODE_NAME_CASE(VFCMADDCSH_RND)
   NODE_NAME_CASE(VPMADD52H)
   NODE_NAME_CASE(VPMADD52L)
   NODE_NAME_CASE(VRNDSCALE)
   NODE_NAME_CASE(STRICT_VRNDSCALE)
   NODE_NAME_CASE(VRNDSCALE_SAE)
   NODE_NAME_CASE(VRNDSCALES)
   NODE_NAME_CASE(VRNDSCALES_SAE)
   NODE_NAME_CASE(VREDUCE)
if (Subtarget.hasSSSE3() && (VT == MVT::v8i16 || VT == MVT::v4i32 ||
                             PostShuffleMask)) {
         auto HOpBuilder = [HorizOpcode](SelectionDAG &DAG, const SDLoc &DL,
                                         ArrayRef<SDValue> Ops) {
           return DAG.getNode(HorizOpcode, DL, Ops[0].getValueType(), Ops);
         };
         SDValue HorizBinOp = SplitOpsAndApply(DAG, Subtarget, SDLoc(N), VT,
                                               {LHS, RHS}, HOpBuilder);
         if (!PostShuffleMask.empty())
           HorizBinOp = DAG.getVectorShuffle(VT, SDLoc(HorizBinOp), HorizBinOp,
                                             DAG.getUNDEF(VT), PostShuffleMask);
         return HorizBinOp;
       }
     }
     break;
   }
 
   return SDValue();
 }
 
+//  Try to combine the following nodes
+//  t29: i64 = X86ISD::Wrapper TargetConstantPool:i64
+//    <i32 -2147483648[float -0.000000e+00]> 0
+//  t27: v16i32[v16f32],ch = X86ISD::VBROADCAST_LOAD
+//    <(load 4 from constant-pool)> t0, t29
+//  [t30: v16i32 = bitcast t27]
+//  t6: v16i32 = xor t7, t27[t30]
+//  t11: v16f32 = bitcast t6
+//  t21: v16f32 = X86ISD::VFMULC[X86ISD::VCFMULC] t11, t8
+//  into X86ISD::VFCMULC[X86ISD::VFMULC] if possible:
+//  t22: v16f32 = bitcast t7
+//  t23: v16f32 = X86ISD::VFCMULC[X86ISD::VFMULC] t8, t22
+//  t24: v32f16 = bitcast t23
+static SDValue combineFMulcFCMulc(SDNode *N, SelectionDAG &DAG,
+                                  const X86Subtarget &Subtarget) {
+  EVT VT = N->getValueType(0);
+  SDValue LHS = N->getOperand(0);
+  SDValue RHS = N->getOperand(1);
+  int CombineOpcode =
+      N->getOpcode() == X86ISD::VFCMULC ? X86ISD::VFMULC : X86ISD::VFCMULC;
+  auto isConjugationConstant = [](const Constant *c) {
+    if (const auto *CI = dyn_cast<ConstantInt>(c)) {
+      APInt ConjugationInt32 = APInt(32, 0x80000000, true);
+      APInt ConjugationInt64 = APInt(64, 0x8000000080000000ULL, true);
+      switch (CI->getBitWidth()) {
+      case 16:
+        return false;
+      case 32:
+        return CI->getValue() == ConjugationInt32;
+      case 64:
+        return CI->getValue() == ConjugationInt64;
+      default:
+        llvm_unreachable("Unexpected bit width");
+      }
+    }
+    if (const auto *CF = dyn_cast<ConstantFP>(c))
+      return CF->isNegativeZeroValue();
+    return false;
+  };
+  auto combineConjugation = [&](SDValue &r) {
+    if (LHS->getOpcode() == ISD::BITCAST && RHS.hasOneUse()) {
+      SDValue XOR = LHS.getOperand(0);
+      if (XOR->getOpcode() == ISD::XOR && XOR.hasOneUse()) {
+        SDValue XORRHS = XOR.getOperand(1);
+        if (XORRHS.getOpcode() == ISD::BITCAST && XORRHS.hasOneUse())
+          XORRHS = XORRHS.getOperand(0);
+        if (XORRHS.getOpcode() == X86ISD::VBROADCAST_LOAD &&
+            XORRHS.getOperand(1).getNumOperands()) {
+          ConstantPoolSDNode *CP =
+              dyn_cast<ConstantPoolSDNode>(XORRHS.getOperand(1).getOperand(0));
+          if (CP && isConjugationConstant(CP->getConstVal())) {
+            SelectionDAG::FlagInserter FlagsInserter(DAG, N);
+            SDValue I2F = DAG.getBitcast(VT, LHS.getOperand(0).getOperand(0));
+            SDValue FCMulC = DAG.getNode(CombineOpcode, SDLoc(N), VT, RHS, I2F);
+            r = DAG.getBitcast(VT, FCMulC);
+            return true;
+          }
+        }
+      }
+    }
+    return false;
+  };
+  SDValue Res;
+  if (combineConjugation(Res))
+    return Res;
+  std::swap(LHS, RHS);
+  if (combineConjugation(Res))
+    return Res;
+  return Res;
+}
+
+//  Try to combine the following nodes
+//  t21: v16f32 = X86ISD::VFMULC/VFCMULC t7, t8
+//  t15: v32f16 = bitcast t21
+//  t16: v32f16 = fadd nnan ninf nsz arcp contract afn reassoc t15, t2
+//  into X86ISD::VFMADDC/VFCMADDC if possible:
+//  t22: v16f32 = bitcast t2
+//  t23: v16f32 = nnan ninf nsz arcp contract afn reassoc
+//                X86ISD::VFMADDC/VFCMADDC t7, t8, t22
+//  t24: v32f16 = bitcast t23
+static SDValue combineFaddCFmul(SDNode *N, SelectionDAG &DAG,
+                                const X86Subtarget &Subtarget) {
+  auto AllowContract = [&DAG](SDNode *N) {
+    return DAG.getTarget().Options.AllowFPOpFusion == FPOpFusion::Fast ||
+           N->getFlags().hasAllowContract();
+  };
+  if (N->getOpcode() != ISD::FADD || !Subtarget.hasFP16() || !AllowContract(N))
+    return SDValue();
+
+  EVT VT = N->getValueType(0);
+  if (VT != MVT::v8f16 && VT != MVT::v16f16 && VT != MVT::v32f16)
+    return SDValue();
+
+  SDValue LHS = N->getOperand(0);
+  SDValue RHS = N->getOperand(1);
+  SDValue CFmul, FAddOp1;
+  auto GetCFmulFrom = [&CFmul, &AllowContract](SDValue N) -> bool {
+    if (!N.hasOneUse() || N.getOpcode() != ISD::BITCAST)
+        return false;
+    SDValue Op0 = N.getOperand(0);
+    unsigned Opcode = Op0.getOpcode();
+    if (Op0.hasOneUse() && AllowContract(Op0.getNode()) &&
+        (Opcode == X86ISD::VFMULC || Opcode == X86ISD::VFCMULC))
+      CFmul = Op0;
+    return !!CFmul;
+  };
+
+  if (GetCFmulFrom(LHS))
+    FAddOp1 = RHS;
+  else if (GetCFmulFrom(RHS))
+    FAddOp1 = LHS;
+  else
+    return SDValue();
+
+  MVT CVT = MVT::getVectorVT(MVT::f32, VT.getVectorNumElements() / 2);
+  assert(CFmul->getValueType(0) == CVT && "Complex type mismatch");
+  FAddOp1 = DAG.getBitcast(CVT, FAddOp1);
+  unsigned newOp = CFmul.getOpcode() == X86ISD::VFMULC ? X86ISD::VFMADDC
+                                                       : X86ISD::VFCMADDC;
+  // FIXME: How do we handle when fast math flags of FADD are different from
+  // CFMUL's?
+  CFmul = DAG.getNode(newOp, SDLoc(N), CVT, FAddOp1, CFmul.getOperand(0),
+                      CFmul.getOperand(1), N->getFlags());
+  return DAG.getBitcast(VT, CFmul);
+}
+
 /// Do target-specific dag combines on floating-point adds/subs.
 static SDValue combineFaddFsub(SDNode *N, SelectionDAG &DAG,
                                const X86Subtarget &Subtarget) {
   if (SDValue HOp = combineToHorizontalAddSub(N, DAG, Subtarget))
     return HOp;
+
+  if (SDValue COp = combineFaddCFmul(N, DAG, Subtarget))
+    return COp;
+
   return SDValue();
 }
 
 /// Attempt to pre-truncate inputs to arithmetic ops if it will simplify
 /// the codegen.
 /// e.g. TRUNC( BINOP( X, Y ) ) --> BINOP( TRUNC( X ), TRUNC( Y ) )
 /// TODO: This overlaps with the generic combiner's visitTRUNCATE. Remove
 ///       anything that is guaranteed to be transformed by DAGCombiner.
SDValue X86TargetLowering::PerformDAGCombine(SDNode *N,
   case ISD::SINT_TO_FP:
   case ISD::STRICT_SINT_TO_FP:
     return combineSIntToFP(N, DAG, DCI, Subtarget);
   case ISD::UINT_TO_FP:
   case ISD::STRICT_UINT_TO_FP:
     return combineUIntToFP(N, DAG, Subtarget);
   case ISD::FADD:
   case ISD::FSUB:           return combineFaddFsub(N, DAG, Subtarget);
+  case X86ISD::VFCMULC:
+  case X86ISD::VFMULC:      return combineFMulcFCMulc(N, DAG, Subtarget);
   case ISD::FNEG:           return combineFneg(N, DAG, DCI, Subtarget);
   case ISD::TRUNCATE:       return combineTruncate(N, DAG, Subtarget);
   case X86ISD::VTRUNC:      return combineVTRUNC(N, DAG, DCI);
   case X86ISD::ANDNP:       return combineAndnp(N, DAG, DCI, Subtarget);
   case X86ISD::FAND:        return combineFAnd(N, DAG, Subtarget);
   case X86ISD::FANDN:       return combineFAndn(N, DAG, Subtarget);
   case X86ISD::FXOR:
   case X86ISD::FOR:         return combineFOr(N, DAG, DCI, Subtarget);

llvm/lib/Target/X86/X86InstrAVX512.td

multiclass AVX512_maskable_custom<bits<8> O, Format F,
                                   string OpcodeStr,
                                   string AttSrcAsm, string IntelSrcAsm,
                                   list<dag> Pattern,
                                   list<dag> MaskingPattern,
                                   list<dag> ZeroMaskingPattern,
                                   string MaskingConstraint = "",
                                   bit IsCommutable = 0,
                                   bit IsKCommutable = 0,
-                                  bit IsKZCommutable = IsCommutable> {
-  let isCommutable = IsCommutable in
+                                  bit IsKZCommutable = IsCommutable,
+                                  string ClobberConstraint = ""> {
+  let isCommutable = IsCommutable, Constraints = ClobberConstraint in
     def NAME: AVX512<O, F, Outs, Ins,
                        OpcodeStr#"\t{"#AttSrcAsm#", $dst|"#
                                      "$dst, "#IntelSrcAsm#"}",
                        Pattern>;
 
   // Prefer over VMOV*rrk Pat<>
   let isCommutable = IsKCommutable in
     def NAME#k: AVX512<O, F, Outs, MaskingIns,
                        OpcodeStr#"\t{"#AttSrcAsm#", $dst {${mask}}|"#
                                      "$dst {${mask}}, "#IntelSrcAsm#"}",
                        MaskingPattern>,
               EVEX_K {
       // In case of the 3src subclass this is overridden with a let.
-      string Constraints = MaskingConstraint;
+      string Constraints = !if(!eq(ClobberConstraint, ""), MaskingConstraint,
+                               !if(!eq(MaskingConstraint, ""), ClobberConstraint,
+                                   !strconcat(ClobberConstraint, ", ", MaskingConstraint)));
     }
 
   // Zero mask does not add any restrictions to commute operands transformation.
   // So, it is Ok to use IsCommutable instead of IsKCommutable.
-  let isCommutable = IsKZCommutable in // Prefer over VMOV*rrkz Pat<>
+  let isCommutable = IsKZCommutable, // Prefer over VMOV*rrkz Pat<>
+      Constraints = ClobberConstraint in
     def NAME#kz: AVX512<O, F, Outs, ZeroMaskingIns,
                        OpcodeStr#"\t{"#AttSrcAsm#", $dst {${mask}} {z}|"#
                                      "$dst {${mask}} {z}, "#IntelSrcAsm#"}",
                        ZeroMaskingPattern>,
               EVEX_KZ;
 }
 
 
 // Common base class of AVX512_maskable and AVX512_maskable_3src.
 multiclass AVX512_maskable_common<bits<8> O, Format F, X86VectorVTInfo _,
                                   dag Outs,
                                   dag Ins, dag MaskingIns, dag ZeroMaskingIns,
                                   string OpcodeStr,
                                   string AttSrcAsm, string IntelSrcAsm,
                                   dag RHS, dag MaskingRHS,
                                   SDPatternOperator Select = vselect_mask,
                                   string MaskingConstraint = "",
                                   bit IsCommutable = 0,
                                   bit IsKCommutable = 0,
-                                  bit IsKZCommutable = IsCommutable> :
+                                  bit IsKZCommutable = IsCommutable,
+                                  string ClobberConstraint = ""> :
   AVX512_maskable_custom<O, F, Outs, Ins, MaskingIns, ZeroMaskingIns, OpcodeStr,
                          AttSrcAsm, IntelSrcAsm,
                          [(set _.RC:$dst, RHS)],
                          [(set _.RC:$dst, MaskingRHS)],
                          [(set _.RC:$dst,
                                (Select _.KRCWM:$mask, RHS, _.ImmAllZerosV))],
                          MaskingConstraint, IsCommutable,
-                         IsKCommutable, IsKZCommutable>;
+                         IsKCommutable, IsKZCommutable, ClobberConstraint>;
 
 // This multiclass generates the unconditional/non-masking, the masking and
 // the zero-masking variant of the vector instruction.  In the masking case, the
 // preserved vector elements come from a new dummy input operand tied to $dst.
 // This version uses a separate dag for non-masking and masking.
 multiclass AVX512_maskable_split<bits<8> O, Format F, X86VectorVTInfo _,
                            dag Outs, dag Ins, string OpcodeStr,
                            string AttSrcAsm, string IntelSrcAsm,
                            dag RHS, dag MaskRHS,
+                           string ClobberConstraint = "",
                            bit IsCommutable = 0, bit IsKCommutable = 0,
                            bit IsKZCommutable = IsCommutable> :
    AVX512_maskable_custom<O, F, Outs, Ins,
                           !con((ins _.RC:$src0, _.KRCWM:$mask), Ins),
                           !con((ins _.KRCWM:$mask), Ins),
                           OpcodeStr, AttSrcAsm, IntelSrcAsm,
                           [(set _.RC:$dst, RHS)],
                           [(set _.RC:$dst,
                               (vselect_mask _.KRCWM:$mask, MaskRHS, _.RC:$src0))],
                           [(set _.RC:$dst,
                               (vselect_mask _.KRCWM:$mask, MaskRHS, _.ImmAllZerosV))],
                           "$src0 = $dst", IsCommutable, IsKCommutable,
-                          IsKZCommutable>;
+                          IsKZCommutable, ClobberConstraint>;
 
 // This multiclass generates the unconditional/non-masking, the masking and
 // the zero-masking variant of the vector instruction.  In the masking case, the
 // preserved vector elements come from a new dummy input operand tied to $dst.
 multiclass AVX512_maskable<bits<8> O, Format F, X86VectorVTInfo _,
                            dag Outs, dag Ins, string OpcodeStr,
                            string AttSrcAsm, string IntelSrcAsm,
                            dag RHS,
                            bit IsCommutable = 0, bit IsKCommutable = 0,
                            bit IsKZCommutable = IsCommutable,
-                           SDPatternOperator Select = vselect_mask> :
+                           SDPatternOperator Select = vselect_mask,
+                           string ClobberConstraint = ""> :
    AVX512_maskable_common<O, F, _, Outs, Ins,
                           !con((ins _.RC:$src0, _.KRCWM:$mask), Ins),
                           !con((ins _.KRCWM:$mask), Ins),
                           OpcodeStr, AttSrcAsm, IntelSrcAsm, RHS,
                           (Select _.KRCWM:$mask, RHS, _.RC:$src0),
                           Select, "$src0 = $dst", IsCommutable, IsKCommutable,
-                          IsKZCommutable>;
+                          IsKZCommutable, ClobberConstraint>;
 
 // This multiclass generates the unconditional/non-masking, the masking and
 // the zero-masking variant of the scalar instruction.
 multiclass AVX512_maskable_scalar<bits<8> O, Format F, X86VectorVTInfo _,
                            dag Outs, dag Ins, string OpcodeStr,
                            string AttSrcAsm, string IntelSrcAsm,
                            dag RHS> :
    AVX512_maskable<O, F, _, Outs, Ins, OpcodeStr, AttSrcAsm, IntelSrcAsm,
defm VMAXCSHZ : avx512_comutable_binop_s<0x5F, "vmaxsh", f16x_info, X86fmaxc,
                                          SchedWriteFCmp.Scl, "VMAXCSH">, T_MAP5XS,
                                          EVEX_4V, VEX_LIG, EVEX_CD8<16, CD8VT1>, SIMD_EXC,
                                          NotEVEX2VEXConvertible;
 
 multiclass avx512_fp_packed<bits<8> opc, string OpcodeStr, SDPatternOperator OpNode,
                             SDPatternOperator MaskOpNode,
                             X86VectorVTInfo _, X86FoldableSchedWrite sched,
                             bit IsCommutable,
-                            bit IsKCommutable = IsCommutable> {
+                            bit IsKCommutable = IsCommutable,
+                            string suffix = _.Suffix,
+                            string ClobberConstraint = "",
+                            bit MayRaiseFPException = 1> {
   let ExeDomain = _.ExeDomain, hasSideEffects = 0,
-      Uses = [MXCSR], mayRaiseFPException = 1 in {
+      Uses = [MXCSR], mayRaiseFPException = MayRaiseFPException in {
   defm rr: AVX512_maskable_split<opc, MRMSrcReg, _, (outs _.RC:$dst),
-                  (ins _.RC:$src1, _.RC:$src2), OpcodeStr#_.Suffix,
+                                 (ins _.RC:$src1, _.RC:$src2), OpcodeStr#suffix,
                                  "$src2, $src1", "$src1, $src2",
                                  (_.VT (OpNode _.RC:$src1, _.RC:$src2)),
-                  (_.VT (MaskOpNode _.RC:$src1, _.RC:$src2)), IsCommutable,
-                  IsKCommutable, IsKCommutable>,
-                  EVEX_4V, Sched<[sched]>;
+                                 (_.VT (MaskOpNode _.RC:$src1, _.RC:$src2)), ClobberConstraint,
+                                 IsCommutable, IsKCommutable, IsKCommutable>, EVEX_4V, Sched<[sched]>;
   let mayLoad = 1 in {
     defm rm: AVX512_maskable_split<opc, MRMSrcMem, _, (outs _.RC:$dst),
-                    (ins _.RC:$src1, _.MemOp:$src2), OpcodeStr#_.Suffix,
+                                   (ins _.RC:$src1, _.MemOp:$src2), OpcodeStr#suffix,
                                    "$src2, $src1", "$src1, $src2",
                                    (OpNode _.RC:$src1, (_.LdFrag addr:$src2)),
-                    (MaskOpNode _.RC:$src1, (_.LdFrag addr:$src2))>,
-                    EVEX_4V, Sched<[sched.Folded, sched.ReadAfterFold]>;
+                                   (MaskOpNode _.RC:$src1, (_.LdFrag addr:$src2)),
+                                   ClobberConstraint>, EVEX_4V, Sched<[sched.Folded, sched.ReadAfterFold]>;
     defm rmb: AVX512_maskable_split<opc, MRMSrcMem, _, (outs _.RC:$dst),
-                     (ins _.RC:$src1, _.ScalarMemOp:$src2), OpcodeStr#_.Suffix,
+                                    (ins _.RC:$src1, _.ScalarMemOp:$src2), OpcodeStr#suffix,
                                     "${src2}"#_.BroadcastStr#", $src1",
                                     "$src1, ${src2}"#_.BroadcastStr,
                                     (OpNode  _.RC:$src1, (_.VT (_.BroadcastLdFrag addr:$src2))),
-                     (MaskOpNode  _.RC:$src1, (_.VT (_.BroadcastLdFrag addr:$src2)))>,
-                     EVEX_4V, EVEX_B,
-                     Sched<[sched.Folded, sched.ReadAfterFold]>;
+                                    (MaskOpNode  _.RC:$src1, (_.VT (_.BroadcastLdFrag addr:$src2))),
+                                    ClobberConstraint>, EVEX_4V, EVEX_B, Sched<[sched.Folded, sched.ReadAfterFold]>;
     }
   }
 }
 
 multiclass avx512_fp_round_packed<bits<8> opc, string OpcodeStr,
                                   SDPatternOperator OpNodeRnd,
-                                  X86FoldableSchedWrite sched, X86VectorVTInfo _> {
+                                  X86FoldableSchedWrite sched, X86VectorVTInfo _,
+                                  string suffix = _.Suffix,
+                                  string ClobberConstraint = ""> {
   let ExeDomain = _.ExeDomain, Uses = [MXCSR] in
   defm rrb: AVX512_maskable<opc, MRMSrcReg, _, (outs _.RC:$dst),
-                  (ins _.RC:$src1, _.RC:$src2, AVX512RC:$rc), OpcodeStr#_.Suffix,
+                  (ins _.RC:$src1, _.RC:$src2, AVX512RC:$rc), OpcodeStr#suffix,
                   "$rc, $src2, $src1", "$src1, $src2, $rc",
-                  (_.VT (OpNodeRnd _.RC:$src1, _.RC:$src2, (i32 timm:$rc)))>,
+                  (_.VT (OpNodeRnd _.RC:$src1, _.RC:$src2, (i32 timm:$rc))),
+                  0, 0, 0, vselect_mask, ClobberConstraint>,
                   EVEX_4V, EVEX_B, EVEX_RC, Sched<[sched]>;
 }
 
 multiclass avx512_fp_sae_packed<bits<8> opc, string OpcodeStr,
                                 SDPatternOperator OpNodeSAE,
                                 X86FoldableSchedWrite sched, X86VectorVTInfo _> {
   let ExeDomain = _.ExeDomain, Uses = [MXCSR] in
   defm rrb: AVX512_maskable<opc, MRMSrcReg, _, (outs _.RC:$dst),
def : Pat<(v8f16 (X86any_VUintToFP (v2i64 (X86VBroadcastld64 addr:$src)))),
             (VCVTUQQ2PHZ128rmb addr:$src)>;
   def : Pat<(X86VMUintToFP (v2i64 (X86VBroadcastld64 addr:$src)),
                            (v8f16 VR128X:$src0), VK2WM:$mask),
             (VCVTUQQ2PHZ128rmbk VR128X:$src0, VK2WM:$mask, addr:$src)>;
   def : Pat<(X86VMUintToFP (v2i64 (X86VBroadcastld64 addr:$src)),
                            v8f16x_info.ImmAllZerosV, VK2WM:$mask),
             (VCVTUQQ2PHZ128rmbkz VK2WM:$mask, addr:$src)>;
 }
+
+let Constraints = "@earlyclobber $dst, $src1 = $dst" in {
+  multiclass avx512_cfmop_rm<bits<8> opc, string OpcodeStr, SDNode OpNode, X86VectorVTInfo _> {
+    defm r: AVX512_maskable_3src<opc, MRMSrcReg, _, (outs _.RC:$dst),
+            (ins _.RC:$src2, _.RC:$src3),
+            OpcodeStr, "$src3, $src2", "$src2, $src3",
+            (_.VT (OpNode _.RC:$src1, _.RC:$src2, _.RC:$src3))>, EVEX_4V;
+
+    defm m: AVX512_maskable_3src<opc, MRMSrcMem, _, (outs _.RC:$dst),
+            (ins _.RC:$src2, _.MemOp:$src3),
+            OpcodeStr, "$src3, $src2", "$src2, $src3",
+            (_.VT (OpNode _.RC:$src1, _.RC:$src2, (_.LdFrag addr:$src3)))>, EVEX_4V;
+
+    defm mb: AVX512_maskable_3src<opc, MRMSrcMem, _, (outs _.RC:$dst),
+            (ins _.RC:$src2, _.ScalarMemOp:$src3),
+            OpcodeStr, !strconcat("${src3}", _.BroadcastStr,", $src2"), !strconcat("$src2, ${src3}", _.BroadcastStr),
+            (_.VT (OpNode _.RC:$src1, _.RC:$src2, (_.VT (_.BroadcastLdFrag addr:$src3))))>, EVEX_B, EVEX_4V;
+  }
+} // Constraints = "@earlyclobber $dst, $src1 = $dst"
+
+multiclass avx512_cfmop_round<bits<8> opc, string OpcodeStr, SDNode OpNode,
+                                 X86VectorVTInfo _> {
+  let Constraints = "@earlyclobber $dst, $src1 = $dst" in
+  defm rb: AVX512_maskable_3src<opc, MRMSrcReg, _, (outs _.RC:$dst),
+          (ins _.RC:$src2, _.RC:$src3, AVX512RC:$rc),
+          OpcodeStr, "$rc, $src3, $src2", "$src2, $src3, $rc",
+          (_.VT (OpNode _.RC:$src1, _.RC:$src2, _.RC:$src3, (i32 timm:$rc)))>,
+          EVEX_4V, EVEX_B, EVEX_RC;
+}
+
+
+multiclass avx512_cfmaop_common<bits<8> opc, string OpcodeStr, SDNode OpNode, SDNode OpNodeRnd> {
+  let Predicates = [HasFP16] in {
+    defm Z    : avx512_cfmop_rm<opc, OpcodeStr, OpNode, v16f32_info>,
+                avx512_cfmop_round<opc, OpcodeStr, OpNodeRnd, v16f32_info>,
+                      EVEX_V512, Sched<[WriteFMAZ]>;
+  }
+  let Predicates = [HasVLX, HasFP16] in {
+    defm Z256 : avx512_cfmop_rm<opc, OpcodeStr, OpNode, v8f32x_info>, EVEX_V256, Sched<[WriteFMAY]>;
+    defm Z128 : avx512_cfmop_rm<opc, OpcodeStr, OpNode, v4f32x_info>, EVEX_V128, Sched<[WriteFMAX]>;
+  }
+}
+
+multiclass avx512_cfmulop_common<bits<8> opc, string OpcodeStr, SDNode OpNode,
+                                 SDNode MaskOpNode, SDNode OpNodeRnd,
+                                 X86SchedWriteWidths sched = SchedWriteFMA> {
+  let Predicates = [HasFP16] in {
+    defm Z    : avx512_fp_packed<opc, OpcodeStr, OpNode, MaskOpNode, v16f32_info,
+                                 sched.ZMM, 0, 0, "", "@earlyclobber $dst", 0>,
+                avx512_fp_round_packed<opc, OpcodeStr, OpNodeRnd, sched.ZMM, v16f32_info,
+                                       "", "@earlyclobber $dst">, EVEX_V512;
+  }
+  let Predicates = [HasVLX, HasFP16] in {
+    defm Z256 : avx512_fp_packed<opc, OpcodeStr, OpNode, MaskOpNode, v8f32x_info,
+                                 sched.YMM, 0, 0, "", "@earlyclobber $dst", 0>, EVEX_V256;
+    defm Z128 : avx512_fp_packed<opc, OpcodeStr, OpNode, MaskOpNode, v4f32x_info,
+                                 sched.XMM, 0, 0, "", "@earlyclobber $dst", 0>, EVEX_V128;
+  }
+}
+
+
+let Uses = [MXCSR] in {
+  defm VFMADDCPH  : avx512_cfmaop_common<0x56, "vfmaddcph", x86vfmaddc, x86vfmaddcRnd>,
+                                    T_MAP6XS, EVEX_CD8<32, CD8VF>;
+  defm VFCMADDCPH : avx512_cfmaop_common<0x56, "vfcmaddcph", x86vfcmaddc, x86vfcmaddcRnd>,
+                                    T_MAP6XD, EVEX_CD8<32, CD8VF>;
+
+  defm VFMULCPH  : avx512_cfmulop_common<0xD6, "vfmulcph", x86vfmulc, x86vfmulc,
+                                         x86vfmulcRnd>, T_MAP6XS, EVEX_CD8<32, CD8VF>;
+  defm VFCMULCPH : avx512_cfmulop_common<0xD6, "vfcmulcph", x86vfcmulc,
+                                         x86vfcmulc, x86vfcmulcRnd>, T_MAP6XD, EVEX_CD8<32, CD8VF>;
+}
+
+
+multiclass avx512_cfmop_sh_common<bits<8> opc, string OpcodeStr, SDNode OpNode, SDNode OpNodeRnd,
+                                  X86SchedWriteWidths sched = SchedWriteFMA> {
+  let Predicates = [HasFP16], Constraints = "@earlyclobber $dst, $src1 = $dst" in {
+    defm r : AVX512_maskable_3src<opc, MRMSrcReg, v4f32x_info, (outs VR128X:$dst),
+                        (ins VR128X:$src2, VR128X:$src3), OpcodeStr,
+                        "$src3, $src2", "$src2, $src3",
+                        (v4f32 (OpNode VR128X:$src1, VR128X:$src2, VR128X:$src3))>,
+                        Sched<[sched.XMM]>;
+    defm m : AVX512_maskable_3src<opc, MRMSrcMem, v4f32x_info, (outs VR128X:$dst),
+                        (ins VR128X:$src2, ssmem:$src3), OpcodeStr,
+                        "$src3, $src2", "$src2, $src3",
+                        (v4f32 (OpNode VR128X:$src1, VR128X:$src2, (sse_load_f32 addr:$src3)))>,
+                        Sched<[sched.XMM.Folded, sched.XMM.ReadAfterFold]>;
+    defm rb : AVX512_maskable_3src<opc,  MRMSrcReg, v4f32x_info, (outs VR128X:$dst),
+                        (ins VR128X:$src2, VR128X:$src3, AVX512RC:$rc), OpcodeStr,
+                        "$rc, $src3, $src2", "$src2, $src3, $rc",
+                        (v4f32 (OpNodeRnd VR128X:$src1, VR128X:$src2, VR128X:$src3, (i32 timm:$rc)))>,
+                        EVEX_B, EVEX_RC, Sched<[sched.XMM]>;
+  }
+}
+
+multiclass avx512_cfmbinop_sh_common<bits<8> opc, string OpcodeStr, SDNode OpNode,
+                                     SDNode OpNodeRnd, X86SchedWriteWidths sched = SchedWriteFMA> {
+  let Predicates = [HasFP16] in {
+    defm rr : AVX512_maskable<opc, MRMSrcReg, f32x_info, (outs VR128X:$dst),
+                        (ins VR128X:$src1, VR128X:$src2), OpcodeStr,
+                        "$src2, $src1", "$src1, $src2",
+                        (v4f32 (OpNode VR128X:$src1, VR128X:$src2)),
+                        0, 0, 0, X86selects, "@earlyclobber $dst">, Sched<[sched.XMM]>;
+    defm rm : AVX512_maskable<opc, MRMSrcMem, f32x_info, (outs VR128X:$dst),
+                        (ins VR128X:$src1, ssmem:$src2), OpcodeStr,
+                        "$src2, $src1", "$src1, $src2",
+                        (v4f32 (OpNode VR128X:$src1, (sse_load_f32 addr:$src2))),
+                        0, 0, 0, X86selects, "@earlyclobber $dst">,
+                        Sched<[sched.XMM.Folded, sched.XMM.ReadAfterFold]>;
+    defm rrb : AVX512_maskable<opc, MRMSrcReg, f32x_info, (outs VR128X:$dst),
+                        (ins VR128X:$src1, VR128X:$src2, AVX512RC:$rc), OpcodeStr,
+                        "$rc, $src2, $src1", "$src1, $src2, $rc",
+                        (OpNodeRnd (v4f32 VR128X:$src1), (v4f32 VR128X:$src2), (i32 timm:$rc)),
+                        0, 0, 0, X86selects, "@earlyclobber $dst">,
+                        EVEX_B, EVEX_RC, Sched<[sched.XMM]>;
+  }
+}
+
+let Uses = [MXCSR] in {
+  defm VFMADDCSHZ  : avx512_cfmop_sh_common<0x57, "vfmaddcsh", x86vfmaddcSh, x86vfmaddcShRnd>,
+                                    T_MAP6XS, EVEX_CD8<32, CD8VT1>, EVEX_V128, EVEX_4V;
+  defm VFCMADDCSHZ : avx512_cfmop_sh_common<0x57, "vfcmaddcsh", x86vfcmaddcSh, x86vfcmaddcShRnd>,
+                                    T_MAP6XD, EVEX_CD8<32, CD8VT1>, EVEX_V128, EVEX_4V;
+
+  defm VFMULCSHZ  : avx512_cfmbinop_sh_common<0xD7, "vfmulcsh", x86vfmulcSh, x86vfmulcShRnd>,
+                                    T_MAP6XS, EVEX_CD8<32, CD8VT1>, EVEX_V128, VEX_LIG, EVEX_4V;
+  defm VFCMULCSHZ : avx512_cfmbinop_sh_common<0xD7, "vfcmulcsh", x86vfcmulcSh, x86vfcmulcShRnd>,
+                                    T_MAP6XD, EVEX_CD8<32, CD8VT1>, EVEX_V128, VEX_LIG, EVEX_4V;
+}

llvm/lib/Target/X86/X86InstrFoldTables.cpp

static const X86MemoryFoldTableEntry MemoryFoldTable2[] = {
   { X86::VEXP2PDZrkz,              X86::VEXP2PDZmkz,              0 },
   { X86::VEXP2PSZrkz,              X86::VEXP2PSZmkz,              0 },
   { X86::VEXPANDPDZ128rrkz,        X86::VEXPANDPDZ128rmkz,        TB_NO_REVERSE },
   { X86::VEXPANDPDZ256rrkz,        X86::VEXPANDPDZ256rmkz,        TB_NO_REVERSE },
   { X86::VEXPANDPDZrrkz,           X86::VEXPANDPDZrmkz,           TB_NO_REVERSE },
   { X86::VEXPANDPSZ128rrkz,        X86::VEXPANDPSZ128rmkz,        TB_NO_REVERSE },
   { X86::VEXPANDPSZ256rrkz,        X86::VEXPANDPSZ256rmkz,        TB_NO_REVERSE },
   { X86::VEXPANDPSZrrkz,           X86::VEXPANDPSZrmkz,           TB_NO_REVERSE },
+  { X86::VFCMULCPHZ128rr,          X86::VFCMULCPHZ128rm,          0 },
+  { X86::VFCMULCPHZ256rr,          X86::VFCMULCPHZ256rm,          0 },
+  { X86::VFCMULCPHZrr,             X86::VFCMULCPHZrm,             0 },
+  { X86::VFCMULCSHZrr,             X86::VFCMULCSHZrm,             TB_NO_REVERSE },
   { X86::VFMADDPD4Yrr,             X86::VFMADDPD4Ymr,             0 },
   { X86::VFMADDPD4rr,              X86::VFMADDPD4mr,              0 },
   { X86::VFMADDPS4Yrr,             X86::VFMADDPS4Ymr,             0 },
   { X86::VFMADDPS4rr,              X86::VFMADDPS4mr,              0 },
   { X86::VFMADDSD4rr,              X86::VFMADDSD4mr,              0 },
   { X86::VFMADDSD4rr_Int,          X86::VFMADDSD4mr_Int,          TB_NO_REVERSE },
   { X86::VFMADDSS4rr,              X86::VFMADDSS4mr,              0 },
   { X86::VFMADDSS4rr_Int,          X86::VFMADDSS4mr_Int,          TB_NO_REVERSE },
   { X86::VFMADDSUBPD4Yrr,          X86::VFMADDSUBPD4Ymr,          0 },
   { X86::VFMADDSUBPD4rr,           X86::VFMADDSUBPD4mr,           0 },
   { X86::VFMADDSUBPS4Yrr,          X86::VFMADDSUBPS4Ymr,          0 },
   { X86::VFMADDSUBPS4rr,           X86::VFMADDSUBPS4mr,           0 },
   { X86::VFMSUBADDPD4Yrr,          X86::VFMSUBADDPD4Ymr,          0 },
   { X86::VFMSUBADDPD4rr,           X86::VFMSUBADDPD4mr,           0 },
   { X86::VFMSUBADDPS4Yrr,          X86::VFMSUBADDPS4Ymr,          0 },
   { X86::VFMSUBADDPS4rr,           X86::VFMSUBADDPS4mr,           0 },
   { X86::VFMSUBPD4Yrr,             X86::VFMSUBPD4Ymr,             0 },
   { X86::VFMSUBPD4rr,              X86::VFMSUBPD4mr,              0 },
   { X86::VFMSUBPS4Yrr,             X86::VFMSUBPS4Ymr,             0 },
   { X86::VFMSUBPS4rr,              X86::VFMSUBPS4mr,              0 },
   { X86::VFMSUBSD4rr,              X86::VFMSUBSD4mr,              0 },
   { X86::VFMSUBSD4rr_Int,          X86::VFMSUBSD4mr_Int,          TB_NO_REVERSE },
   { X86::VFMSUBSS4rr,              X86::VFMSUBSS4mr,              0 },
   { X86::VFMSUBSS4rr_Int,          X86::VFMSUBSS4mr_Int,          TB_NO_REVERSE },
+  { X86::VFMULCPHZ128rr,           X86::VFMULCPHZ128rm,           0 },
+  { X86::VFMULCPHZ256rr,           X86::VFMULCPHZ256rm,           0 },
+  { X86::VFMULCPHZrr,              X86::VFMULCPHZrm,              0 },
+  { X86::VFMULCSHZrr,              X86::VFMULCSHZrm,              TB_NO_REVERSE },
   { X86::VFNMADDPD4Yrr,            X86::VFNMADDPD4Ymr,            0 },
   { X86::VFNMADDPD4rr,             X86::VFNMADDPD4mr,             0 },
   { X86::VFNMADDPS4Yrr,            X86::VFNMADDPS4Ymr,            0 },
   { X86::VFNMADDPS4rr,             X86::VFNMADDPS4mr,             0 },
   { X86::VFNMADDSD4rr,             X86::VFNMADDSD4mr,             0 },
   { X86::VFNMADDSD4rr_Int,         X86::VFNMADDSD4mr_Int,         TB_NO_REVERSE },
   { X86::VFNMADDSS4rr,             X86::VFNMADDSS4mr,             0 },
   { X86::VFNMADDSS4rr_Int,         X86::VFNMADDSS4mr_Int,         TB_NO_REVERSE },
static const X86MemoryFoldTableEntry MemoryFoldTable3[] = {
   { X86::VEXP2PDZrk,                 X86::VEXP2PDZmk,                 0 },
   { X86::VEXP2PSZrk,                 X86::VEXP2PSZmk,                 0 },
   { X86::VEXPANDPDZ128rrk,           X86::VEXPANDPDZ128rmk,           TB_NO_REVERSE },
   { X86::VEXPANDPDZ256rrk,           X86::VEXPANDPDZ256rmk,           TB_NO_REVERSE },
   { X86::VEXPANDPDZrrk,              X86::VEXPANDPDZrmk,              TB_NO_REVERSE },
   { X86::VEXPANDPSZ128rrk,           X86::VEXPANDPSZ128rmk,           TB_NO_REVERSE },
   { X86::VEXPANDPSZ256rrk,           X86::VEXPANDPSZ256rmk,           TB_NO_REVERSE },
   { X86::VEXPANDPSZrrk,              X86::VEXPANDPSZrmk,              TB_NO_REVERSE },
+  { X86::VFCMADDCPHZ128r,            X86::VFCMADDCPHZ128m,            0 },
+  { X86::VFCMADDCPHZ256r,            X86::VFCMADDCPHZ256m,            0 },
+  { X86::VFCMADDCPHZr,               X86::VFCMADDCPHZm,               0 },
+  { X86::VFCMADDCSHZr,               X86::VFCMADDCSHZm,               TB_NO_REVERSE },
+  { X86::VFCMULCPHZ128rrkz,          X86::VFCMULCPHZ128rmkz,          0 },
+  { X86::VFCMULCPHZ256rrkz,          X86::VFCMULCPHZ256rmkz,          0 },
+  { X86::VFCMULCPHZrrkz,             X86::VFCMULCPHZrmkz,             0 },
+  { X86::VFCMULCSHZrrkz,             X86::VFCMULCSHZrmkz,             TB_NO_REVERSE },
   { X86::VFIXUPIMMPDZ128rri,         X86::VFIXUPIMMPDZ128rmi,         0 },
   { X86::VFIXUPIMMPDZ256rri,         X86::VFIXUPIMMPDZ256rmi,         0 },
   { X86::VFIXUPIMMPDZrri,            X86::VFIXUPIMMPDZrmi,            0 },
   { X86::VFIXUPIMMPSZ128rri,         X86::VFIXUPIMMPSZ128rmi,         0 },
   { X86::VFIXUPIMMPSZ256rri,         X86::VFIXUPIMMPSZ256rmi,         0 },
   { X86::VFIXUPIMMPSZrri,            X86::VFIXUPIMMPSZrmi,            0 },
   { X86::VFIXUPIMMSDZrri,            X86::VFIXUPIMMSDZrmi,            TB_NO_REVERSE },
   { X86::VFIXUPIMMSSZrri,            X86::VFIXUPIMMSSZrmi,            TB_NO_REVERSE },
static const X86MemoryFoldTableEntry MemoryFoldTable3[] = {
   { X86::VFMADD231SDr,               X86::VFMADD231SDm,               0 },
   { X86::VFMADD231SDr_Int,           X86::VFMADD231SDm_Int,           TB_NO_REVERSE },
   { X86::VFMADD231SHZr,              X86::VFMADD231SHZm,              0 },
   { X86::VFMADD231SHZr_Int,          X86::VFMADD231SHZm_Int,          TB_NO_REVERSE },
   { X86::VFMADD231SSZr,              X86::VFMADD231SSZm,              0 },
   { X86::VFMADD231SSZr_Int,          X86::VFMADD231SSZm_Int,          TB_NO_REVERSE },
   { X86::VFMADD231SSr,               X86::VFMADD231SSm,               0 },
   { X86::VFMADD231SSr_Int,           X86::VFMADD231SSm_Int,           TB_NO_REVERSE },
+  { X86::VFMADDCPHZ128r,             X86::VFMADDCPHZ128m,             0 },
+  { X86::VFMADDCPHZ256r,             X86::VFMADDCPHZ256m,             0 },
+  { X86::VFMADDCPHZr,                X86::VFMADDCPHZm,                0 },
+  { X86::VFMADDCSHZr,                X86::VFMADDCSHZm,                TB_NO_REVERSE },
   { X86::VFMADDPD4Yrr,               X86::VFMADDPD4Yrm,               0 },
   { X86::VFMADDPD4rr,                X86::VFMADDPD4rm,                0 },
   { X86::VFMADDPS4Yrr,               X86::VFMADDPS4Yrm,               0 },
   { X86::VFMADDPS4rr,                X86::VFMADDPS4rm,                0 },
   { X86::VFMADDSD4rr,                X86::VFMADDSD4rm,                0 },
   { X86::VFMADDSD4rr_Int,            X86::VFMADDSD4rm_Int,            TB_NO_REVERSE },
   { X86::VFMADDSS4rr,                X86::VFMADDSS4rm,                0 },
   { X86::VFMADDSS4rr_Int,            X86::VFMADDSS4rm_Int,            TB_NO_REVERSE },
static const X86MemoryFoldTableEntry MemoryFoldTable3[] = {
   { X86::VFMSUBPD4Yrr,               X86::VFMSUBPD4Yrm,               0 },
   { X86::VFMSUBPD4rr,                X86::VFMSUBPD4rm,                0 },
   { X86::VFMSUBPS4Yrr,               X86::VFMSUBPS4Yrm,               0 },
   { X86::VFMSUBPS4rr,                X86::VFMSUBPS4rm,                0 },
   { X86::VFMSUBSD4rr,                X86::VFMSUBSD4rm,                0 },
   { X86::VFMSUBSD4rr_Int,            X86::VFMSUBSD4rm_Int,            TB_NO_REVERSE },
   { X86::VFMSUBSS4rr,                X86::VFMSUBSS4rm,                0 },
   { X86::VFMSUBSS4rr_Int,            X86::VFMSUBSS4rm_Int,            TB_NO_REVERSE },
+  { X86::VFMULCPHZ128rrkz,           X86::VFMULCPHZ128rmkz,           0 },
+  { X86::VFMULCPHZ256rrkz,           X86::VFMULCPHZ256rmkz,           0 },
+  { X86::VFMULCPHZrrkz,              X86::VFMULCPHZrmkz,              0 },
+  { X86::VFMULCSHZrrkz,              X86::VFMULCSHZrmkz,              TB_NO_REVERSE },
   { X86::VFNMADD132PDYr,             X86::VFNMADD132PDYm,             0 },
   { X86::VFNMADD132PDZ128r,          X86::VFNMADD132PDZ128m,          0 },
   { X86::VFNMADD132PDZ256r,          X86::VFNMADD132PDZ256m,          0 },
   { X86::VFNMADD132PDZr,             X86::VFNMADD132PDZm,             0 },
   { X86::VFNMADD132PDr,              X86::VFNMADD132PDm,              0 },
   { X86::VFNMADD132PHZ128r,          X86::VFNMADD132PHZ128m,          0 },
   { X86::VFNMADD132PHZ256r,          X86::VFNMADD132PHZ256m,          0 },
   { X86::VFNMADD132PHZr,             X86::VFNMADD132PHZm,             0 },
static const X86MemoryFoldTableEntry MemoryFoldTable4[] = {
   { X86::VDIVSHZrr_Intk,            X86::VDIVSHZrm_Intk,            TB_NO_REVERSE },
   { X86::VDIVSSZrr_Intk,            X86::VDIVSSZrm_Intk,            TB_NO_REVERSE },
   { X86::VDPBF16PSZ128rk,           X86::VDPBF16PSZ128mk,           0 },
   { X86::VDPBF16PSZ128rkz,          X86::VDPBF16PSZ128mkz,          0 },
   { X86::VDPBF16PSZ256rk,           X86::VDPBF16PSZ256mk,           0 },
   { X86::VDPBF16PSZ256rkz,          X86::VDPBF16PSZ256mkz,          0 },
   { X86::VDPBF16PSZrk,              X86::VDPBF16PSZmk,              0 },
   { X86::VDPBF16PSZrkz,             X86::VDPBF16PSZmkz,             0 },
+  { X86::VFCMADDCPHZ128rk,          X86::VFCMADDCPHZ128mk,          0 },
+  { X86::VFCMADDCPHZ128rkz,         X86::VFCMADDCPHZ128mkz,         0 },
+  { X86::VFCMADDCPHZ256rk,          X86::VFCMADDCPHZ256mk,          0 },
+  { X86::VFCMADDCPHZ256rkz,         X86::VFCMADDCPHZ256mkz,         0 },
+  { X86::VFCMADDCPHZrk,             X86::VFCMADDCPHZmk,             0 },
+  { X86::VFCMADDCPHZrkz,            X86::VFCMADDCPHZmkz,            0 },
+  { X86::VFCMADDCSHZrk,             X86::VFCMADDCSHZmk,             TB_NO_REVERSE },
+  { X86::VFCMADDCSHZrkz,            X86::VFCMADDCSHZmkz,            TB_NO_REVERSE },
+  { X86::VFCMULCPHZ128rrk,          X86::VFCMULCPHZ128rmk,          0 },
+  { X86::VFCMULCPHZ256rrk,          X86::VFCMULCPHZ256rmk,          0 },
+  { X86::VFCMULCPHZrrk,             X86::VFCMULCPHZrmk,             0 },
+  { X86::VFCMULCSHZrrk,             X86::VFCMULCSHZrmk,             TB_NO_REVERSE },
   { X86::VFIXUPIMMPDZ128rrik,       X86::VFIXUPIMMPDZ128rmik,       0 },
   { X86::VFIXUPIMMPDZ128rrikz,      X86::VFIXUPIMMPDZ128rmikz,      0 },
   { X86::VFIXUPIMMPDZ256rrik,       X86::VFIXUPIMMPDZ256rmik,       0 },
   { X86::VFIXUPIMMPDZ256rrikz,      X86::VFIXUPIMMPDZ256rmikz,      0 },
   { X86::VFIXUPIMMPDZrrik,          X86::VFIXUPIMMPDZrmik,          0 },
   { X86::VFIXUPIMMPDZrrikz,         X86::VFIXUPIMMPDZrmikz,         0 },
   { X86::VFIXUPIMMPSZ128rrik,       X86::VFIXUPIMMPSZ128rmik,       0 },
   { X86::VFIXUPIMMPSZ128rrikz,      X86::VFIXUPIMMPSZ128rmikz,      0 },
static const X86MemoryFoldTableEntry MemoryFoldTable4[] = {
   { X86::VFMADD231PSZrk,            X86::VFMADD231PSZmk,            0 },
   { X86::VFMADD231PSZrkz,           X86::VFMADD231PSZmkz,           0 },
   { X86::VFMADD231SDZr_Intk,        X86::VFMADD231SDZm_Intk,        TB_NO_REVERSE },
   { X86::VFMADD231SDZr_Intkz,       X86::VFMADD231SDZm_Intkz,       TB_NO_REVERSE },
   { X86::VFMADD231SHZr_Intk,        X86::VFMADD231SHZm_Intk,        TB_NO_REVERSE },
   { X86::VFMADD231SHZr_Intkz,       X86::VFMADD231SHZm_Intkz,       TB_NO_REVERSE },
   { X86::VFMADD231SSZr_Intk,        X86::VFMADD231SSZm_Intk,        TB_NO_REVERSE },
   { X86::VFMADD231SSZr_Intkz,       X86::VFMADD231SSZm_Intkz,       TB_NO_REVERSE },
+  { X86::VFMADDCPHZ128rk,           X86::VFMADDCPHZ128mk,           0 },
+  { X86::VFMADDCPHZ128rkz,          X86::VFMADDCPHZ128mkz,          0 },
+  { X86::VFMADDCPHZ256rk,           X86::VFMADDCPHZ256mk,           0 },
+  { X86::VFMADDCPHZ256rkz,          X86::VFMADDCPHZ256mkz,          0 },
+  { X86::VFMADDCPHZrk,              X86::VFMADDCPHZmk,              0 },
+  { X86::VFMADDCPHZrkz,             X86::VFMADDCPHZmkz,             0 },
+  { X86::VFMADDCSHZrk,              X86::VFMADDCSHZmk,              TB_NO_REVERSE },
+  { X86::VFMADDCSHZrkz,             X86::VFMADDCSHZmkz,             TB_NO_REVERSE },
   { X86::VFMADDSUB132PDZ128rk,      X86::VFMADDSUB132PDZ128mk,      0 },
   { X86::VFMADDSUB132PDZ128rkz,     X86::VFMADDSUB132PDZ128mkz,     0 },
   { X86::VFMADDSUB132PDZ256rk,      X86::VFMADDSUB132PDZ256mk,      0 },
   { X86::VFMADDSUB132PDZ256rkz,     X86::VFMADDSUB132PDZ256mkz,     0 },
   { X86::VFMADDSUB132PDZrk,         X86::VFMADDSUB132PDZmk,         0 },
   { X86::VFMADDSUB132PDZrkz,        X86::VFMADDSUB132PDZmkz,        0 },
   { X86::VFMADDSUB132PHZ128rk,      X86::VFMADDSUB132PHZ128mk,      0 },
   { X86::VFMADDSUB132PHZ128rkz,     X86::VFMADDSUB132PHZ128mkz,     0 },
static const X86MemoryFoldTableEntry MemoryFoldTable4[] = {
   { X86::VFMSUBADD231PHZrk,         X86::VFMSUBADD231PHZmk,         0 },
   { X86::VFMSUBADD231PHZrkz,        X86::VFMSUBADD231PHZmkz,        0 },
   { X86::VFMSUBADD231PSZ128rk,      X86::VFMSUBADD231PSZ128mk,      0 },
   { X86::VFMSUBADD231PSZ128rkz,     X86::VFMSUBADD231PSZ128mkz,     0 },
   { X86::VFMSUBADD231PSZ256rk,      X86::VFMSUBADD231PSZ256mk,      0 },
   { X86::VFMSUBADD231PSZ256rkz,     X86::VFMSUBADD231PSZ256mkz,     0 },
   { X86::VFMSUBADD231PSZrk,         X86::VFMSUBADD231PSZmk,         0 },
   { X86::VFMSUBADD231PSZrkz,        X86::VFMSUBADD231PSZmkz,        0 },
+  { X86::VFMULCPHZ128rrk,           X86::VFMULCPHZ128rmk,           0 },
+  { X86::VFMULCPHZ256rrk,           X86::VFMULCPHZ256rmk,           0 },
+  { X86::VFMULCPHZrrk,              X86::VFMULCPHZrmk,              0 },
+  { X86::VFMULCSHZrrk,              X86::VFMULCSHZrmk,              TB_NO_REVERSE },
   { X86::VFNMADD132PDZ128rk,        X86::VFNMADD132PDZ128mk,        0 },
   { X86::VFNMADD132PDZ128rkz,       X86::VFNMADD132PDZ128mkz,       0 },
   { X86::VFNMADD132PDZ256rk,        X86::VFNMADD132PDZ256mk,        0 },
   { X86::VFNMADD132PDZ256rkz,       X86::VFNMADD132PDZ256mkz,       0 },
   { X86::VFNMADD132PDZrk,           X86::VFNMADD132PDZmk,           0 },
   { X86::VFNMADD132PDZrkz,          X86::VFNMADD132PDZmkz,          0 },
   { X86::VFNMADD132PHZ128rk,        X86::VFNMADD132PHZ128mk,        0 },
   { X86::VFNMADD132PHZ128rkz,       X86::VFNMADD132PHZ128mkz,       0 },

llvm/lib/Target/X86/X86InstrFragmentsSIMD.td

def X86vp2intersect : SDNode<"X86ISD::VP2INTERSECT",
                               SDTypeProfile<1, 2, [SDTCisVT<0, untyped>,
                                                    SDTCisVec<1>, SDTCisSameAs<1, 2>]>>;
 
 def SDTIFma : SDTypeProfile<1, 3, [SDTCisInt<0>, SDTCisSameAs<0,1>,
                            SDTCisSameAs<1,2>, SDTCisSameAs<1,3>]>;
 def x86vpmadd52l     : SDNode<"X86ISD::VPMADD52L",     SDTIFma, [SDNPCommutative]>;
 def x86vpmadd52h     : SDNode<"X86ISD::VPMADD52H",     SDTIFma, [SDNPCommutative]>;
 
+def x86vfmaddc       : SDNode<"X86ISD::VFMADDC",       SDTFPTernaryOp>;
+def x86vfmaddcRnd    : SDNode<"X86ISD::VFMADDC_RND",   SDTFmaRound>;
+def x86vfcmaddc      : SDNode<"X86ISD::VFCMADDC",      SDTFPTernaryOp>;
+def x86vfcmaddcRnd   : SDNode<"X86ISD::VFCMADDC_RND",  SDTFmaRound>;
+def x86vfmulc        : SDNode<"X86ISD::VFMULC",        SDTFPBinOp>;
+def x86vfmulcRnd     : SDNode<"X86ISD::VFMULC_RND",    SDTFPBinOpRound>;
+def x86vfcmulc       : SDNode<"X86ISD::VFCMULC",       SDTFPBinOp>;
+def x86vfcmulcRnd    : SDNode<"X86ISD::VFCMULC_RND",   SDTFPBinOpRound>;
+
+def x86vfmaddcSh     : SDNode<"X86ISD::VFMADDCSH",     SDTFPTernaryOp>;
+def x86vfcmaddcSh    : SDNode<"X86ISD::VFCMADDCSH",    SDTFPTernaryOp>;
+def x86vfmulcSh      : SDNode<"X86ISD::VFMULCSH",      SDTFPBinOp>;
+def x86vfcmulcSh     : SDNode<"X86ISD::VFCMULCSH",     SDTFPBinOp>;
+def x86vfmaddcShRnd  : SDNode<"X86ISD::VFMADDCSH_RND", SDTFmaRound>;
+def x86vfcmaddcShRnd : SDNode<"X86ISD::VFCMADDCSH_RND",SDTFmaRound>;
+def x86vfmulcShRnd   : SDNode<"X86ISD::VFMULCSH_RND",  SDTFPBinOpRound>;
+def x86vfcmulcShRnd  : SDNode<"X86ISD::VFCMULCSH_RND", SDTFPBinOpRound>;
+
 def X86rsqrt14   : SDNode<"X86ISD::RSQRT14",  SDTFPUnaryOp>;
 def X86rcp14     : SDNode<"X86ISD::RCP14",    SDTFPUnaryOp>;
 
 // VNNI
 def SDTVnni : SDTypeProfile<1, 3, [SDTCisVec<0>, SDTCisSameAs<0,1>,
                                    SDTCisSameAs<1,2>, SDTCisSameAs<1,3>]>;
 def X86Vpdpbusd  : SDNode<"X86ISD::VPDPBUSD", SDTVnni>;
 def X86Vpdpbusds : SDNode<"X86ISD::VPDPBUSDS", SDTVnni>;

llvm/lib/Target/X86/X86IntrinsicsInfo.h

 
 namespace llvm {
 
 enum IntrinsicType : uint16_t {
   CVTNEPS2BF16_MASK,
   GATHER, SCATTER, PREFETCH, RDSEED, RDRAND, RDPMC, RDTSC, XTEST, XGETBV, ADX, FPCLASSS,
   INTR_TYPE_1OP, INTR_TYPE_2OP, INTR_TYPE_3OP, INTR_TYPE_4OP_IMM8,
   INTR_TYPE_3OP_IMM8,
+  FMA_OP_MASK, FMA_OP_MASKZ,
   CMP_MASK_CC,CMP_MASK_SCALAR_CC, VSHIFT, COMI, COMI_RM, BLENDV, BEXTRI,
   CVTPD2PS_MASK,
   INTR_TYPE_1OP_SAE, INTR_TYPE_2OP_SAE,
   INTR_TYPE_1OP_MASK_SAE, INTR_TYPE_2OP_MASK_SAE, INTR_TYPE_3OP_MASK_SAE,
   INTR_TYPE_1OP_MASK, INTR_TYPE_2OP_MASK,
   IFMA_OP, VPERM_2OP, INTR_TYPE_SCALAR_MASK, INTR_TYPE_SCALAR_MASK_SAE,
   INTR_TYPE_SCALAR_MASK_RND,
   INTR_TYPE_3OP_SCALAR_MASK_SAE,
static const IntrinsicData  IntrinsicsWithoutChain[] = {
   X86_INTRINSIC_DATA(avx512fp16_mask_vcvttph2w_512, INTR_TYPE_1OP_MASK_SAE,
                      X86ISD::CVTTP2SI, X86ISD::CVTTP2SI_SAE),
   X86_INTRINSIC_DATA(avx512fp16_mask_vcvtudq2ph_128, TRUNCATE_TO_REG,
                      X86ISD::CVTUI2P, X86ISD::MCVTUI2P),
   X86_INTRINSIC_DATA(avx512fp16_mask_vcvtuqq2ph_128, TRUNCATE_TO_REG,
                      X86ISD::CVTUI2P, X86ISD::MCVTUI2P),
   X86_INTRINSIC_DATA(avx512fp16_mask_vcvtuqq2ph_256, TRUNCATE_TO_REG,
                      X86ISD::CVTUI2P, X86ISD::MCVTUI2P),
+  X86_INTRINSIC_DATA(avx512fp16_mask_vfcmadd_cph_128, FMA_OP_MASK,  X86ISD::VFCMADDC,  0),
+  X86_INTRINSIC_DATA(avx512fp16_mask_vfcmadd_cph_256, FMA_OP_MASK,  X86ISD::VFCMADDC,  0),
+  X86_INTRINSIC_DATA(avx512fp16_mask_vfcmadd_cph_512, FMA_OP_MASK,  X86ISD::VFCMADDC,  X86ISD::VFCMADDC_RND),
+  X86_INTRINSIC_DATA(avx512fp16_mask_vfcmadd_csh, FMA_OP_MASK,  X86ISD::VFCMADDCSH, X86ISD::VFCMADDCSH_RND),
+  X86_INTRINSIC_DATA(avx512fp16_mask_vfcmul_cph_128, INTR_TYPE_2OP_MASK,  X86ISD::VFCMULC,  0),
+  X86_INTRINSIC_DATA(avx512fp16_mask_vfcmul_cph_256, INTR_TYPE_2OP_MASK,  X86ISD::VFCMULC,  0),
+  X86_INTRINSIC_DATA(avx512fp16_mask_vfcmul_cph_512, INTR_TYPE_2OP_MASK,  X86ISD::VFCMULC,  X86ISD::VFCMULC_RND),
+  X86_INTRINSIC_DATA(avx512fp16_mask_vfcmul_csh, INTR_TYPE_SCALAR_MASK,  X86ISD::VFCMULCSH, X86ISD::VFCMULCSH_RND),
+  X86_INTRINSIC_DATA(avx512fp16_mask_vfmadd_cph_128, FMA_OP_MASK,  X86ISD::VFMADDC,  0),
+  X86_INTRINSIC_DATA(avx512fp16_mask_vfmadd_cph_256, FMA_OP_MASK,  X86ISD::VFMADDC,  0),
+  X86_INTRINSIC_DATA(avx512fp16_mask_vfmadd_cph_512, FMA_OP_MASK,  X86ISD::VFMADDC,  X86ISD::VFMADDC_RND),
+  X86_INTRINSIC_DATA(avx512fp16_mask_vfmadd_csh, FMA_OP_MASK,  X86ISD::VFMADDCSH, X86ISD::VFMADDCSH_RND),
+  X86_INTRINSIC_DATA(avx512fp16_mask_vfmul_cph_128, INTR_TYPE_2OP_MASK,  X86ISD::VFMULC,  0),
+  X86_INTRINSIC_DATA(avx512fp16_mask_vfmul_cph_256, INTR_TYPE_2OP_MASK,  X86ISD::VFMULC,  0),
+  X86_INTRINSIC_DATA(avx512fp16_mask_vfmul_cph_512, INTR_TYPE_2OP_MASK,  X86ISD::VFMULC,  X86ISD::VFMULC_RND),
+  X86_INTRINSIC_DATA(avx512fp16_mask_vfmul_csh, INTR_TYPE_SCALAR_MASK,  X86ISD::VFMULCSH, X86ISD::VFMULCSH_RND),
+  X86_INTRINSIC_DATA(avx512fp16_maskz_vfcmadd_cph_128, FMA_OP_MASKZ,  X86ISD::VFCMADDC,  0),
+  X86_INTRINSIC_DATA(avx512fp16_maskz_vfcmadd_cph_256, FMA_OP_MASKZ,  X86ISD::VFCMADDC,  0),
+  X86_INTRINSIC_DATA(avx512fp16_maskz_vfcmadd_cph_512, FMA_OP_MASKZ,  X86ISD::VFCMADDC,  X86ISD::VFCMADDC_RND),
+  X86_INTRINSIC_DATA(avx512fp16_maskz_vfcmadd_csh, FMA_OP_MASKZ,  X86ISD::VFCMADDCSH, X86ISD::VFCMADDCSH_RND),
+  X86_INTRINSIC_DATA(avx512fp16_maskz_vfmadd_cph_128, FMA_OP_MASKZ,  X86ISD::VFMADDC,  0),
+  X86_INTRINSIC_DATA(avx512fp16_maskz_vfmadd_cph_256, FMA_OP_MASKZ,  X86ISD::VFMADDC,  0),
+  X86_INTRINSIC_DATA(avx512fp16_maskz_vfmadd_cph_512, FMA_OP_MASKZ,  X86ISD::VFMADDC,  X86ISD::VFMADDC_RND),
+  X86_INTRINSIC_DATA(avx512fp16_maskz_vfmadd_csh, FMA_OP_MASKZ,  X86ISD::VFMADDCSH, X86ISD::VFMADDCSH_RND),
   X86_INTRINSIC_DATA(avx512fp16_max_ph_128, INTR_TYPE_2OP, X86ISD::FMAX, 0),
   X86_INTRINSIC_DATA(avx512fp16_max_ph_256, INTR_TYPE_2OP, X86ISD::FMAX, 0),
   X86_INTRINSIC_DATA(avx512fp16_max_ph_512, INTR_TYPE_2OP_SAE, X86ISD::FMAX, X86ISD::FMAX_SAE),
   X86_INTRINSIC_DATA(avx512fp16_min_ph_128, INTR_TYPE_2OP, X86ISD::FMIN, 0),
   X86_INTRINSIC_DATA(avx512fp16_min_ph_256, INTR_TYPE_2OP, X86ISD::FMIN, 0),
   X86_INTRINSIC_DATA(avx512fp16_min_ph_512, INTR_TYPE_2OP_SAE, X86ISD::FMIN, X86ISD::FMIN_SAE),
   X86_INTRINSIC_DATA(avx512fp16_mul_ph_512, INTR_TYPE_2OP, ISD::FMUL, X86ISD::FMUL_RND),
   X86_INTRINSIC_DATA(avx512fp16_sqrt_ph_512, INTR_TYPE_1OP, ISD::FSQRT, X86ISD::FSQRT_RND),

llvm/test/CodeGen/X86/avx512cfma-intrinsics.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl -mattr=+avx512bw -mattr=+avx512fp16 -mattr=+avx512vl | FileCheck %s
+
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.128(<4 x float>, <4 x float>, <4 x float>, i8)
+declare <4 x float> @llvm.x86.avx512fp16.maskz.vfmadd.cph.128(<4 x float>, <4 x float>, <4 x float>, i8)
+
+define <4 x float> @test_int_x86_avx512fp8_mask_cfmadd_ph_bst(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp8_mask_cfmadd_ph_bst:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmaddcph {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm1, %xmm0 {%k1}
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.128(<4 x float> %x0, <4 x float> %x1, <4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, i8 %x3)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_int_x86_avx512fp8_mask_cfmadd_ph_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp8_mask_cfmadd_ph_128:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmaddcph %xmm2, %xmm1, %xmm0 {%k1}
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_int_x86_avx512fp8_maskz_cfmadd_ph_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp8_maskz_cfmadd_ph_128:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmaddcph %xmm2, %xmm1, %xmm0 {%k1} {z}
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.maskz.vfmadd.cph.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_int_x86_avx512fp8_cfmadd_ph_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2){
+; CHECK-LABEL: test_int_x86_avx512fp8_cfmadd_ph_128:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfmaddcph %xmm0, %xmm1, %xmm2
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.128(<4 x float> %x2, <4 x float> %x1, <4 x float> %x0, i8 -1)
+  ret <4 x float> %res
+}
+
+
+declare <8 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.256(<8 x float>, <8 x float>, <8 x float>, i8)
+declare <8 x float> @llvm.x86.avx512fp16.maskz.vfmadd.cph.256(<8 x float>, <8 x float>, <8 x float>, i8)
+
+define <8 x float> @test_int_x86_avx512fp16_mask_cfmadd_ph_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_mask_cfmadd_ph_256:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmaddcph %ymm2, %ymm1, %ymm0 {%k1}
+; CHECK-NEXT:    retq
+  %res = call <8 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
+  ret <8 x float> %res
+}
+
+define <8 x float> @test_int_x86_avx512fp16_maskz_cfmadd_ph_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_maskz_cfmadd_ph_256:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmaddcph %ymm2, %ymm1, %ymm0 {%k1} {z}
+; CHECK-NEXT:    retq
+  %res = call <8 x float> @llvm.x86.avx512fp16.maskz.vfmadd.cph.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
+  ret <8 x float> %res
+}
+
+define <8 x float> @test_int_x86_avx512fp16_cfmadd_ph_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2){
+; CHECK-LABEL: test_int_x86_avx512fp16_cfmadd_ph_256:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfmaddcph %ymm0, %ymm1, %ymm2
+; CHECK-NEXT:    vmovaps %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.256(<8 x float> %x2, <8 x float> %x1, <8 x float> %x0, i8 -1)
+  ret <8 x float> %res
+}
+
+
+declare <16 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
+declare <16 x float> @llvm.x86.avx512fp16.maskz.vfmadd.cph.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
+
+define <16 x float> @test_int_x86_avx512fp16_mask_cfmadd_ph_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_mask_cfmadd_ph_512:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmaddcph %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT:    retq
+  %res = call <16 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
+  ret <16 x float> %res
+}
+
+define <16 x float> @test_int_x86_avx512fp16_maskz_cfmadd_ph_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_maskz_cfmadd_ph_512:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmaddcph %zmm2, %zmm1, %zmm0 {%k1} {z}
+; CHECK-NEXT:    retq
+  %res = call <16 x float> @llvm.x86.avx512fp16.maskz.vfmadd.cph.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
+  ret <16 x float> %res
+}
+
+define <16 x float> @test_int_x86_avx512fp16_cfmadd_ph_512_rn(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_cfmadd_ph_512_rn:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfmaddcph {rz-sae}, %zmm0, %zmm1, %zmm2
+; CHECK-NEXT:    vmovaps %zmm2, %zmm0
+; CHECK-NEXT:    retq
+  %res = call <16 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.512(<16 x float> %x2, <16 x float> %x1, <16 x float> %x0, i16 -1, i32 11)
+  ret <16 x float> %res
+}
+
+define <16 x float> @test_int_x86_avx512fp16_cfmadd_ph_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_cfmadd_ph_512:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfmaddcph %zmm0, %zmm1, %zmm2
+; CHECK-NEXT:    vmovaps %zmm2, %zmm0
+; CHECK-NEXT:    retq
+  %res = call <16 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.512(<16 x float> %x2, <16 x float> %x1, <16 x float> %x0, i16 -1, i32 4)
+  ret <16 x float> %res
+}
+
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.128(<4 x float>, <4 x float>, <4 x float>, i8)
+declare <4 x float> @llvm.x86.avx512fp16.maskz.vfcmadd.cph.128(<4 x float>, <4 x float>, <4 x float>, i8)
+
+define <4 x float> @test_int_x86_avx512fp8_mask_cfcmadd_ph_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp8_mask_cfcmadd_ph_128:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmaddcph %xmm2, %xmm1, %xmm0 {%k1}
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_int_x86_avx512fp8_maskz_cfcmadd_ph_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp8_maskz_cfcmadd_ph_128:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmaddcph %xmm2, %xmm1, %xmm0 {%k1} {z}
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.maskz.vfcmadd.cph.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_int_x86_avx512fp8_cfcmadd_ph_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2){
+; CHECK-LABEL: test_int_x86_avx512fp8_cfcmadd_ph_128:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfcmaddcph %xmm0, %xmm1, %xmm2
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.128(<4 x float> %x2, <4 x float> %x1, <4 x float> %x0, i8 -1)
+  ret <4 x float> %res
+}
+
+
+declare <8 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.256(<8 x float>, <8 x float>, <8 x float>, i8)
+declare <8 x float> @llvm.x86.avx512fp16.maskz.vfcmadd.cph.256(<8 x float>, <8 x float>, <8 x float>, i8)
+
+define <8 x float> @test_int_x86_avx512fp16_mask_cfcmadd_ph_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_mask_cfcmadd_ph_256:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmaddcph %ymm2, %ymm1, %ymm0 {%k1}
+; CHECK-NEXT:    retq
+  %res = call <8 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
+  ret <8 x float> %res
+}
+
+define <8 x float> @test_int_x86_avx512fp16_maskz_cfcmadd_ph_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_maskz_cfcmadd_ph_256:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmaddcph %ymm2, %ymm1, %ymm0 {%k1} {z}
+; CHECK-NEXT:    retq
+  %res = call <8 x float> @llvm.x86.avx512fp16.maskz.vfcmadd.cph.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
+  ret <8 x float> %res
+}
+
+define <8 x float> @test_int_x86_avx512fp16_cfcmadd_ph_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2){
+; CHECK-LABEL: test_int_x86_avx512fp16_cfcmadd_ph_256:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfcmaddcph %ymm0, %ymm1, %ymm2
+; CHECK-NEXT:    vmovaps %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.256(<8 x float> %x2, <8 x float> %x1, <8 x float> %x0, i8 -1)
+  ret <8 x float> %res
+}
+
+
+declare <16 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
+declare <16 x float> @llvm.x86.avx512fp16.maskz.vfcmadd.cph.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
+
+define <16 x float> @test_int_x86_avx512fp16_mask_cfcmadd_ph_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_mask_cfcmadd_ph_512:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmaddcph %zmm2, %zmm1, %zmm0 {%k1}
+; CHECK-NEXT:    retq
+  %res = call <16 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
+  ret <16 x float> %res
+}
+
+define <16 x float> @test_int_x86_avx512fp16_maskz_cfcmadd_ph_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_maskz_cfcmadd_ph_512:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmaddcph %zmm2, %zmm1, %zmm0 {%k1} {z}
+; CHECK-NEXT:    retq
+  %res = call <16 x float> @llvm.x86.avx512fp16.maskz.vfcmadd.cph.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
+  ret <16 x float> %res
+}
+
+define <16 x float> @test_int_x86_avx512fp16_cfcmadd_ph_512_rn(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_cfcmadd_ph_512_rn:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfcmaddcph {rz-sae}, %zmm0, %zmm1, %zmm2
+; CHECK-NEXT:    vmovaps %zmm2, %zmm0
+; CHECK-NEXT:    retq
+  %res = call <16 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.512(<16 x float> %x2, <16 x float> %x1, <16 x float> %x0, i16 -1, i32 11)
+  ret <16 x float> %res
+}
+
+define <16 x float> @test_int_x86_avx512fp16_cfcmadd_ph_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_cfcmadd_ph_512:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfcmaddcph %zmm0, %zmm1, %zmm2
+; CHECK-NEXT:    vmovaps %zmm2, %zmm0
+; CHECK-NEXT:    retq
+  %res = call <16 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.512(<16 x float> %x2, <16 x float> %x1, <16 x float> %x0, i16 -1, i32 4)
+  ret <16 x float> %res
+}

llvm/test/CodeGen/X86/avx512cfmul-intrinsics.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl -mattr=+avx512bw -mattr=+avx512fp16 -mattr=+avx512vl | FileCheck %s
+
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float>, <4 x float>, <4 x float>, i8)
+
+define <4 x float> @test_int_x86_avx512fp8_mask_cfmul_ph_bst(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp8_mask_cfmul_ph_bst:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmulcph {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float> %x0, <4 x float> <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>, <4 x float> %x2, i8 %x3)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_int_x86_avx512fp8_mask_cfmul_ph_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp8_mask_cfmul_ph_128:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmulcph %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_int_x86_avx512fp8_maskz_cfmul_ph_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp8_maskz_cfmul_ph_128:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmulcph %xmm1, %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float> %x0, <4 x float> %x1, <4 x float> zeroinitializer, i8 %x3)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_int_x86_avx512fp8_cfmul_ph_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2){
+; CHECK-LABEL: test_int_x86_avx512fp8_cfmul_ph_128:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfmulcph %xmm1, %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float> %x2, <4 x float> %x1, <4 x float> %x0, i8 -1)
+  ret <4 x float> %res
+}
+
+declare <8 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.256(<8 x float>, <8 x float>, <8 x float>, i8)
+
+define <8 x float> @test_int_x86_avx512fp16_mask_cfmul_ph_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_mask_cfmul_ph_256:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmulcph %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    vmovaps %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
+  ret <8 x float> %res
+}
+
+define <8 x float> @test_int_x86_avx512fp16_maskz_cfmul_ph_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_maskz_cfmul_ph_256:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmulcph %ymm1, %ymm0, %ymm2 {%k1} {z}
+; CHECK-NEXT:    vmovaps %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.256(<8 x float> %x0, <8 x float> %x1, <8 x float> zeroinitializer, i8 %x3)
+  ret <8 x float> %res
+}
+
+define <8 x float> @test_int_x86_avx512fp16_cfmul_ph_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2){
+; CHECK-LABEL: test_int_x86_avx512fp16_cfmul_ph_256:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfmulcph %ymm1, %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.256(<8 x float> %x2, <8 x float> %x1, <8 x float> %x0, i8 -1)
+  ret <8 x float> %res
+}
+
+declare <16 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
+
+define <16 x float> @test_int_x86_avx512fp16_mask_cfmul_ph_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_mask_cfmul_ph_512:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmulcph %zmm1, %zmm0, %zmm2 {%k1}
+; CHECK-NEXT:    vmovaps %zmm2, %zmm0
+; CHECK-NEXT:    retq
+  %res = call <16 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
+  ret <16 x float> %res
+}
+
+define <16 x float> @test_int_x86_avx512fp16_maskz_cfmul_ph_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_maskz_cfmul_ph_512:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmulcph %zmm1, %zmm0, %zmm2 {%k1} {z}
+; CHECK-NEXT:    vmovaps %zmm2, %zmm0
+; CHECK-NEXT:    retq
+  %res = call <16 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.512(<16 x float> %x0, <16 x float> %x1, <16 x float> zeroinitializer, i16 %x3, i32 4)
+  ret <16 x float> %res
+}
+
+define <16 x float> @test_int_x86_avx512fp16_cfmul_ph_512_rn(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_cfmul_ph_512_rn:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfmulcph {rz-sae}, %zmm1, %zmm2, %zmm0
+; CHECK-NEXT:    retq
+  %res = call <16 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.512(<16 x float> %x2, <16 x float> %x1, <16 x float> %x0, i16 -1, i32 11)
+  ret <16 x float> %res
+}
+
+define <16 x float> @test_int_x86_avx512fp16_cfmul_ph_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_cfmul_ph_512:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfmulcph %zmm1, %zmm2, %zmm0
+; CHECK-NEXT:    retq
+  %res = call <16 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.512(<16 x float> %x2, <16 x float> %x1, <16 x float> %x0, i16 -1, i32 4)
+  ret <16 x float> %res
+}
+
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.128(<4 x float>, <4 x float>, <4 x float>, i8)
+
+define <4 x float> @test_int_x86_avx512fp8_mask_cfcmul_ph_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp8_mask_cfcmul_ph_128:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmulcph %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_int_x86_avx512fp8_maskz_cfcmul_ph_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp8_maskz_cfcmul_ph_128:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmulcph %xmm1, %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.128(<4 x float> %x0, <4 x float> %x1, <4 x float> zeroinitializer, i8 %x3)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_int_x86_avx512fp8_cfcmul_ph_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2){
+; CHECK-LABEL: test_int_x86_avx512fp8_cfcmul_ph_128:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfcmulcph %xmm1, %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.128(<4 x float> %x2, <4 x float> %x1, <4 x float> %x0, i8 -1)
+  ret <4 x float> %res
+}
+
+declare <8 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.256(<8 x float>, <8 x float>, <8 x float>, i8)
+
+define <8 x float> @test_int_x86_avx512fp16_mask_cfcmul_ph_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_mask_cfcmul_ph_256:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmulcph %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-NEXT:    vmovaps %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
+  ret <8 x float> %res
+}
+
+define <8 x float> @test_int_x86_avx512fp16_maskz_cfcmul_ph_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_maskz_cfcmul_ph_256:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmulcph %ymm1, %ymm0, %ymm2 {%k1} {z}
+; CHECK-NEXT:    vmovaps %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.256(<8 x float> %x0, <8 x float> %x1, <8 x float> zeroinitializer, i8 %x3)
+  ret <8 x float> %res
+}
+
+define <8 x float> @test_int_x86_avx512fp16_cfcmul_ph_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2){
+; CHECK-LABEL: test_int_x86_avx512fp16_cfcmul_ph_256:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfcmulcph %ymm1, %ymm2, %ymm0
+; CHECK-NEXT:    retq
+  %res = call <8 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.256(<8 x float> %x2, <8 x float> %x1, <8 x float> %x0, i8 -1)
+  ret <8 x float> %res
+}
+
+declare <16 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
+
+define <16 x float> @test_int_x86_avx512fp16_mask_cfcmul_ph_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_mask_cfcmul_ph_512:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmulcph %zmm1, %zmm0, %zmm2 {%k1}
+; CHECK-NEXT:    vmovaps %zmm2, %zmm0
+; CHECK-NEXT:    retq
+  %res = call <16 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
+  ret <16 x float> %res
+}
+
+define <16 x float> @test_int_x86_avx512fp16_maskz_cfcmul_ph_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_maskz_cfcmul_ph_512:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmulcph %zmm1, %zmm0, %zmm2 {%k1} {z}
+; CHECK-NEXT:    vmovaps %zmm2, %zmm0
+; CHECK-NEXT:    retq
+  %res = call <16 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.512(<16 x float> %x0, <16 x float> %x1, <16 x float> zeroinitializer, i16 %x3, i32 4)
+  ret <16 x float> %res
+}
+
+define <16 x float> @test_int_x86_avx512fp16_cfcmul_ph_512_rn(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_cfcmul_ph_512_rn:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfcmulcph {rz-sae}, %zmm1, %zmm2, %zmm0
+; CHECK-NEXT:    retq
+  %res = call <16 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.512(<16 x float> %x2, <16 x float> %x1, <16 x float> %x0, i16 -1, i32 11)
+  ret <16 x float> %res
+}
+
+define <16 x float> @test_int_x86_avx512fp16_cfcmul_ph_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
+; CHECK-LABEL: test_int_x86_avx512fp16_cfcmul_ph_512:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfcmulcph %zmm1, %zmm2, %zmm0
+; CHECK-NEXT:    retq
+  %res = call <16 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.512(<16 x float> %x2, <16 x float> %x1, <16 x float> %x0, i16 -1, i32 4)
+  ret <16 x float> %res
+}

llvm/test/CodeGen/X86/avx512cfmulsh-instrinsics.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl -mattr=+avx512bw -mattr=+avx512fp16 -mattr=+avx512vl | FileCheck %s
+
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfmul.csh(<4 x float>, <4 x float>, <4 x float>, i8, i32)
+declare <4 x float> @llvm.x86.avx512fp16.maskz.vfmul.csh(<4 x float>, <4 x float>, <4 x float>, i8, i32)
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.csh(<4 x float>, <4 x float>, <4 x float>, i8, i32)
+declare <4 x float> @llvm.x86.avx512fp16.maskz.vfcmul.csh(<4 x float>, <4 x float>, <4 x float>, i8, i32)
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.csh(<4 x float>, <4 x float>, <4 x float>, i8, i32)
+declare <4 x float> @llvm.x86.avx512fp16.maskz.vfmadd.csh(<4 x float>, <4 x float>, <4 x float>, i8, i32)
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.csh(<4 x float>, <4 x float>, <4 x float>, i8, i32)
+declare <4 x float> @llvm.x86.avx512fp16.maskz.vfcmadd.csh(<4 x float>, <4 x float>, <4 x float>, i8, i32)
+
+;; no mask, no rounding
+
+define <4 x float> @test_nm_nr_int_x86_avx512fp16_mask_cfmul_sh(<4 x float> %x0, <4 x float> %x1) {
+; CHECK-LABEL: test_nm_nr_int_x86_avx512fp16_mask_cfmul_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfmulcsh %xmm1, %xmm0, %xmm2
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfmul.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> undef, i8 -1, i32 4)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_nm_nr_int_x86_avx512fp16_mask_cfcmul_sh(<4 x float> %x0, <4 x float> %x1) {
+; CHECK-LABEL: test_nm_nr_int_x86_avx512fp16_mask_cfcmul_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfcmulcsh %xmm1, %xmm0, %xmm2
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> undef, i8 -1, i32 4)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_nm_nr_int_x86_avx512fp16_cfmadd_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2) {
+; CHECK-LABEL: test_nm_nr_int_x86_avx512fp16_cfmadd_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfmaddcsh %xmm2, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1, i32 4)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_nm_nr_int_x86_avx512fp16_cfcmadd_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2) {
+; CHECK-LABEL: test_nm_nr_int_x86_avx512fp16_cfcmadd_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfcmaddcsh %xmm2, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1, i32 4)
+  ret <4 x float> %res
+}
+
+;; no mask, rounding
+
+define <4 x float> @test_nm_r_int_x86_avx512fp16_mask_cfmul_sh(<4 x float> %x0, <4 x float> %x1) {
+; CHECK-LABEL: test_nm_r_int_x86_avx512fp16_mask_cfmul_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfmulcsh {rd-sae}, %xmm1, %xmm0, %xmm2
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfmul.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> undef, i8 -1, i32 9)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_nm_r_int_x86_avx512fp16_mask_cfcmul_sh(<4 x float> %x0, <4 x float> %x1) {
+; CHECK-LABEL: test_nm_r_int_x86_avx512fp16_mask_cfcmul_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfcmulcsh {rd-sae}, %xmm1, %xmm0, %xmm2
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> undef, i8 -1, i32 9)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_nm_r_int_x86_avx512fp16_mask_cfmadd_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2) {
+; CHECK-LABEL: test_nm_r_int_x86_avx512fp16_mask_cfmadd_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfmaddcsh {rd-sae}, %xmm2, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1, i32 9)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_nm_r_int_x86_avx512fp16_mask_cfcmadd_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2) {
+; CHECK-LABEL: test_nm_r_int_x86_avx512fp16_mask_cfcmadd_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    vfcmaddcsh {rd-sae}, %xmm2, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1, i32 9)
+  ret <4 x float> %res
+}
+
+;; mask, no rounding
+
+define <4 x float> @test_m_nr_int_x86_avx512fp16_mask_cfmul_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_m_nr_int_x86_avx512fp16_mask_cfmul_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmulcsh %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfmul.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3, i32 4)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_m_nr_int_x86_avx512fp16_mask_cfcmul_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_m_nr_int_x86_avx512fp16_mask_cfcmul_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmulcsh %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3, i32 4)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_m_nr_int_x86_avx512fp16_mask_cfmadd_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_m_nr_int_x86_avx512fp16_mask_cfmadd_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmaddcsh %xmm2, %xmm1, %xmm0 {%k1}
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3, i32 4)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_m_nr_int_x86_avx512fp16_mask_cfcmadd_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_m_nr_int_x86_avx512fp16_mask_cfcmadd_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmaddcsh %xmm2, %xmm1, %xmm0 {%k1}
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3, i32 4)
+  ret <4 x float> %res
+}
+
+;; mask, rounding
+
+define <4 x float> @test_int_x86_avx512fp16_mask_cfmul_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512fp16_mask_cfmul_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmulcsh {rd-sae}, %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfmul.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3, i32 9)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_int_x86_avx512fp16_mask_cfcmul_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512fp16_mask_cfcmul_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmulcsh {rd-sae}, %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3, i32 9)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_int_x86_avx512fp16_mask_cfmadd_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512fp16_mask_cfmadd_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmaddcsh {rd-sae}, %xmm2, %xmm1, %xmm0 {%k1}
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3, i32 9)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_int_x86_avx512fp16_mask_cfcmadd_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512fp16_mask_cfcmadd_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmaddcsh {rd-sae}, %xmm2, %xmm1, %xmm0 {%k1}
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3, i32 9)
+  ret <4 x float> %res
+}
+
+;; maskz, no rounding
+
+define <4 x float> @test_m_nr_int_x86_avx512fp16_maskz_cfmul_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_m_nr_int_x86_avx512fp16_maskz_cfmul_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmulcsh %xmm1, %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfmul.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> zeroinitializer, i8 %x3, i32 4)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_m_nr_int_x86_avx512fp16_maskz_cfcmul_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_m_nr_int_x86_avx512fp16_maskz_cfcmul_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmulcsh %xmm1, %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> zeroinitializer, i8 %x3, i32 4)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_m_nr_int_x86_avx512fp16_maskz_cfmadd_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_m_nr_int_x86_avx512fp16_maskz_cfmadd_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmaddcsh %xmm2, %xmm1, %xmm0 {%k1} {z}
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.maskz.vfmadd.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3, i32 4)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_m_nr_int_x86_avx512fp16_maskz_cfcmadd_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_m_nr_int_x86_avx512fp16_maskz_cfcmadd_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmaddcsh %xmm2, %xmm1, %xmm0 {%k1} {z}
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.maskz.vfcmadd.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3, i32 4)
+  ret <4 x float> %res
+}
+
+;; maskz, rounding
+
+define <4 x float> @test_int_x86_avx512fp16_maskz_cfmul_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512fp16_maskz_cfmul_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmulcsh {rd-sae}, %xmm1, %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfmul.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> zeroinitializer, i8 %x3, i32 9)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_int_x86_avx512fp16_maskz_cfcmul_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512fp16_maskz_cfcmul_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmulcsh {rd-sae}, %xmm1, %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> zeroinitializer, i8 %x3, i32 9)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_int_x86_avx512fp16_maskz_cfmadd_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512fp16_maskz_cfmadd_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfmaddcsh {rd-sae}, %xmm2, %xmm1, %xmm0 {%k1} {z}
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.maskz.vfmadd.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3, i32 9)
+  ret <4 x float> %res
+}
+
+define <4 x float> @test_int_x86_avx512fp16_maskz_cfcmadd_sh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512fp16_maskz_cfcmadd_sh:
+; CHECK:       ## %bb.0:
+; CHECK-NEXT:    kmovd %edi, %k1
+; CHECK-NEXT:    vfcmaddcsh {rd-sae}, %xmm2, %xmm1, %xmm0 {%k1} {z}
+; CHECK-NEXT:    retq
+  %res = call <4 x float> @llvm.x86.avx512fp16.maskz.vfcmadd.csh(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3, i32 9)
+  ret <4 x float> %res
+}

llvm/test/CodeGen/X86/avx512fp16-combine-vfmulc-fadd.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=avx512fp16 --fp-contract=fast --enable-unsafe-fp-math | FileCheck %s
+
+define dso_local <32 x half> @test1(<32 x half> %acc.coerce, <32 x half> %lhs.coerce, <32 x half> %rhs.coerce) {
+; CHECK-LABEL: test1:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vfmaddcph %zmm2, %zmm1, %zmm0
+; CHECK-NEXT:    retq
+entry:
+  %0 = bitcast <32 x half> %lhs.coerce to <16 x float>
+  %1 = bitcast <32 x half> %rhs.coerce to <16 x float>
+  %2 = tail call fast <16 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.512(<16 x float> %0, <16 x float> %1, <16 x float> zeroinitializer, i16 -1, i32 4)
+  %3 = bitcast <16 x float> %2 to <32 x half>
+  %add.i.i = fadd fast <32 x half> %3, %acc.coerce
+  ret <32 x half> %add.i.i
+}
+
+define dso_local <16 x half> @test2(<16 x half> %acc.coerce, <16 x half> %lhs.coerce, <16 x half> %rhs.coerce) {
+; CHECK-LABEL: test2:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vfmaddcph %ymm2, %ymm1, %ymm0
+; CHECK-NEXT:    retq
+entry:
+  %0 = bitcast <16 x half> %lhs.coerce to <8 x float>
+  %1 = bitcast <16 x half> %rhs.coerce to <8 x float>
+  %2 = tail call fast <8 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.256(<8 x float> %0, <8 x float> %1, <8 x float> zeroinitializer, i8 -1)
+  %3 = bitcast <8 x float> %2 to <16 x half>
+  %add.i.i = fadd fast <16 x half> %3, %acc.coerce
+  ret <16 x half> %add.i.i
+}
+
+define dso_local <8 x half> @test3(<8 x half> %acc.coerce, <8 x half> %lhs.coerce, <8 x half> %rhs.coerce) {
+; CHECK-LABEL: test3:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vfmaddcph %xmm2, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+entry:
+  %0 = bitcast <8 x half> %lhs.coerce to <4 x float>
+  %1 = bitcast <8 x half> %rhs.coerce to <4 x float>
+  %2 = tail call fast <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float> %0, <4 x float> %1, <4 x float> zeroinitializer, i8 -1)
+  %3 = bitcast <4 x float> %2 to <8 x half>
+  %add.i.i = fadd fast <8 x half> %3, %acc.coerce
+  ret <8 x half> %add.i.i
+}
+
+
+define dso_local <8 x half> @test4(<8 x half> %acc.coerce, <8 x half> %lhs.coerce, <8 x half> %rhs.coerce) {
+; CHECK-LABEL: test4:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vfmaddcph %xmm2, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+entry:
+  %0 = bitcast <8 x half> %lhs.coerce to <4 x float>
+  %1 = bitcast <8 x half> %rhs.coerce to <4 x float>
+  %2 = tail call fast <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float> %0, <4 x float> %1, <4 x float> zeroinitializer, i8 -1)
+  %3 = bitcast <4 x float> %2 to <8 x half>
+  %add.i.i = fadd fast <8 x half> %acc.coerce, %3
+  ret <8 x half> %add.i.i
+}
+
+declare <16 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.512(<16 x float>, <16 x float>, <16 x float>, i16, i32 immarg)
+declare <8 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.256(<8 x float>, <8 x float>, <8 x float>, i8)
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float>, <4 x float>, <4 x float>, i8)

llvm/test/CodeGen/X86/avx512fp16-combine-xor-vfmulc-fadd.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=avx512fp16 --fp-contract=fast --enable-unsafe-fp-math | FileCheck %s
+
+define dso_local <32 x half> @test1(<32 x half> %acc.coerce, <32 x half> %lhs.coerce.conj, <32 x half> %rhs.coerce) local_unnamed_addr #0 {
+; CHECK-LABEL: test1:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vfcmaddcph %zmm1, %zmm2, %zmm0
+; CHECK-NEXT:    retq
+entry:
+  %0 = bitcast <32 x half> %lhs.coerce.conj to <16 x i32>
+  %xor.i.i = xor <16 x i32> %0, <i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648>
+  %1 = bitcast <16 x i32> %xor.i.i to <16 x float>
+  %2 = bitcast <32 x half> %rhs.coerce to <16 x float>
+  %3 = tail call fast <16 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.512(<16 x float> %1, <16 x float> %2, <16 x float> zeroinitializer, i16 -1, i32 4) #2
+  %4 = bitcast <16 x float> %3 to <32 x half>
+  %add = fadd fast <32 x half> %4, %acc.coerce
+  ret <32 x half> %add
+}
+
+define dso_local <32 x half> @test2(<32 x half> %acc.coerce, <32 x half> %lhs.coerce.conj, <32 x half> %rhs.coerce) local_unnamed_addr #0 {
+; CHECK-LABEL: test2:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vfcmaddcph %zmm1, %zmm2, %zmm0
+; CHECK-NEXT:    retq
+entry:
+  %0 = bitcast <32 x half> %lhs.coerce.conj to <16 x i32>
+  %xor.i.i = xor <16 x i32> %0, <i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648>
+  %1 = bitcast <16 x i32> %xor.i.i to <16 x float>
+  %2 = bitcast <32 x half> %rhs.coerce to <16 x float>
+  %3 = tail call fast <16 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.512(<16 x float> %2, <16 x float> %1, <16 x float> zeroinitializer, i16 -1, i32 4) #2
+  %4 = bitcast <16 x float> %3 to <32 x half>
+  %add = fadd fast <32 x half> %4, %acc.coerce
+  ret <32 x half> %add
+}
+
+define dso_local <16 x half> @test3(<16 x half> %acc.coerce, <16 x half> %lhs.coerce.conj, <16 x half> %rhs.coerce) local_unnamed_addr #0 {
+; CHECK-LABEL: test3:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vfcmaddcph %ymm1, %ymm2, %ymm0
+; CHECK-NEXT:    retq
+entry:
+  %0 = bitcast <16 x half> %lhs.coerce.conj to <8 x i32>
+  %xor.i.i = xor <8 x i32> %0, <i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648>
+  %1 = bitcast <8 x i32> %xor.i.i to <8 x float>
+  %2 = bitcast <16 x half> %rhs.coerce to <8 x float>
+  %3 = tail call fast <8 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.256(<8 x float> %1, <8 x float> %2, <8 x float> zeroinitializer, i8 -1) #2
+  %4 = bitcast <8 x float> %3 to <16 x half>
+  %add = fadd fast <16 x half> %4, %acc.coerce
+  ret <16 x half> %add
+}
+
+define dso_local <8 x half> @test4(<8 x half> %acc.coerce, <8 x half> %lhs.coerce.conj, <8 x half> %rhs.coerce) local_unnamed_addr #0 {
+; CHECK-LABEL: test4:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vfcmaddcph %xmm1, %xmm2, %xmm0
+; CHECK-NEXT:    retq
+entry:
+  %0 = bitcast <8 x half> %lhs.coerce.conj to <4 x i32>
+  %xor.i.i = xor <4 x i32> %0, <i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648>
+  %1 = bitcast <4 x i32> %xor.i.i to <4 x float>
+  %2 = bitcast <8 x half> %rhs.coerce to <4 x float>
+  %3 = tail call fast <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float> %1, <4 x float> %2, <4 x float> zeroinitializer, i8 -1) #2
+  %4 = bitcast <4 x float> %3 to <8 x half>
+  %add = fadd fast <8 x half> %4, %acc.coerce
+  ret <8 x half> %add
+}
+
+define dso_local <8 x half> @test5(<8 x half> %acc.coerce, <8 x half> %lhs.coerce.conj, <8 x half> %rhs.coerce) local_unnamed_addr #0 {
+; CHECK-LABEL: test5:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vfcmaddcph %xmm1, %xmm2, %xmm0
+; CHECK-NEXT:    retq
+entry:
+  %0 = bitcast <8 x half> %lhs.coerce.conj to <4 x i32>
+  %xor.i.i = xor <4 x i32> <i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648>, %0
+  %1 = bitcast <4 x i32> %xor.i.i to <4 x float>
+  %2 = bitcast <8 x half> %rhs.coerce to <4 x float>
+  %3 = tail call fast <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float> %1, <4 x float> %2, <4 x float> zeroinitializer, i8 -1) #2
+  %4 = bitcast <4 x float> %3 to <8 x half>
+  %add = fadd fast <8 x half> %4, %acc.coerce
+  ret <8 x half> %add
+}
+
+define dso_local <8 x half> @test6(<8 x half> %acc.coerce, <8 x half> %lhs.coerce.conj, <8 x half> %rhs.coerce) local_unnamed_addr #0 {
+; CHECK-LABEL: test6:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vxorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm1, %xmm1
+; CHECK-NEXT:    vfmaddcph %xmm2, %xmm1, %xmm0
+; CHECK-NEXT:    retq
+entry:
+  %0 = bitcast <8 x half> %lhs.coerce.conj to <4 x i32>
+  %xor.i.i = xor <4 x i32> <i32 1, i32 1, i32 1, i32 1>, %0
+  %1 = bitcast <4 x i32> %xor.i.i to <4 x float>
+  %2 = bitcast <8 x half> %rhs.coerce to <4 x float>
+  %3 = tail call fast <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float> %1, <4 x float> %2, <4 x float> zeroinitializer, i8 -1) #2
+  %4 = bitcast <4 x float> %3 to <8 x half>
+  %add = fadd fast <8 x half> %4, %acc.coerce
+  ret <8 x half> %add
+}
+
+define dso_local <8 x half> @test7(<8 x half> %acc.coerce, <8 x half> %lhs.coerce.conj, <8 x half> %rhs.coerce) local_unnamed_addr #0 {
+; CHECK-LABEL: test7:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vfcmaddcph %xmm1, %xmm2, %xmm0
+; CHECK-NEXT:    retq
+entry:
+  %0 = bitcast <8 x half> %lhs.coerce.conj to <4 x i32>
+  %xor.i.i = xor <4 x i32> <i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648>, %0
+  %1 = bitcast <4 x i32> %xor.i.i to <4 x float>
+  %2 = bitcast <8 x half> %rhs.coerce to <4 x float>
+  %3 = tail call fast <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float> %1, <4 x float> %2, <4 x float> zeroinitializer, i8 -1) #2
+  %4 = bitcast <4 x float> %3 to <8 x half>
+  %add = fadd fast <8 x half> %acc.coerce, %4
+  ret <8 x half> %add
+}
+
+define dso_local <8 x half> @test8(<8 x half> %acc.coerce, <4 x float> %lhs.coerce.conj, <8 x half> %rhs.coerce) local_unnamed_addr #0 {
+; CHECK-LABEL: test8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vfcmaddcph %xmm1, %xmm2, %xmm0
+; CHECK-NEXT:    retq
+entry:
+  %0 = bitcast <4 x float> %lhs.coerce.conj to <4 x i32>
+  %xor.i.i = xor <4 x i32> <i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648>, %0
+  %1 = bitcast <4 x i32> %xor.i.i to <4 x float>
+  %2 = bitcast <8 x half> %rhs.coerce to <4 x float>
+  %3 = tail call fast <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float> %1, <4 x float> %2, <4 x float> zeroinitializer, i8 -1) #2
+  %4 = bitcast <4 x float> %3 to <8 x half>
+  %add = fadd fast <8 x half> %acc.coerce, %4
+  ret <8 x half> %add
+}
+
+define dso_local <32 x half> @test9(<32 x half> %acc.coerce, <8 x i64> %lhs.coerce.conj, <32 x half> %rhs.coerce) local_unnamed_addr #0 {
+; CHECK-LABEL: test9:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vfcmaddcph %zmm1, %zmm2, %zmm0
+; CHECK-NEXT:    retq
+entry:
+  %xor1.i = xor <8 x i64> %lhs.coerce.conj, <i64 -9223372034707292160, i64 -9223372034707292160, i64 -9223372034707292160, i64 -9223372034707292160, i64 -9223372034707292160, i64 -9223372034707292160, i64 -9223372034707292160, i64 -9223372034707292160>
+  %0 = bitcast <8 x i64> %xor1.i to <16 x float>
+  %1 = bitcast <32 x half> %rhs.coerce to <16 x float>
+  %2 = tail call fast <16 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.512(<16 x float> %0, <16 x float> %1, <16 x float> zeroinitializer, i16 -1, i32 4) #2
+  %3 = bitcast <16 x float> %2 to <32 x half>
+  %add = fadd fast <32 x half> %3, %acc.coerce
+  ret <32 x half> %add
+}
+
+declare <16 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.512(<16 x float>, <16 x float>, <16 x float>, i16, i32 immarg)
+declare <8 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.256(<8 x float>, <8 x float>, <8 x float>, i8)
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float>, <4 x float>, <4 x float>, i8)

llvm/test/CodeGen/X86/avx512fp16-combine-xor-vfmulc.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=avx512fp16 --fp-contract=fast --enable-unsafe-fp-math | FileCheck %s
+
+define dso_local <32 x half> @test1(<32 x half> %lhs.coerce.conj, <32 x half> %rhs.coerce) local_unnamed_addr #0 {
+; CHECK-LABEL: test1:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vfcmulcph %zmm0, %zmm1, %zmm2
+; CHECK-NEXT:    vmovaps %zmm2, %zmm0
+; CHECK-NEXT:    retq
+entry:
+  %0 = bitcast <32 x half> %lhs.coerce.conj to <16 x i32>
+  %xor.i.i = xor <16 x i32> %0, <i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648>
+  %1 = bitcast <16 x i32> %xor.i.i to <16 x float>
+  %2 = bitcast <32 x half> %rhs.coerce to <16 x float>
+  %3 = tail call fast <16 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.512(<16 x float> %1, <16 x float> %2, <16 x float> zeroinitializer, i16 -1, i32 4) #2
+  %4 = bitcast <16 x float> %3 to <32 x half>
+  ret <32 x half> %4
+}
+
+; Function Attrs: nounwind readnone uwtable
+define dso_local <16 x half> @test2(<16 x half> %lhs.coerce.conj, <16 x half> %rhs.coerce) local_unnamed_addr #0 {
+; CHECK-LABEL: test2:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vfcmulcph %ymm0, %ymm1, %ymm2
+; CHECK-NEXT:    vmovaps %ymm2, %ymm0
+; CHECK-NEXT:    retq
+entry:
+  %0 = bitcast <16 x half> %lhs.coerce.conj to <8 x i32>
+  %xor.i.i = xor <8 x i32> %0, <i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648>
+  %1 = bitcast <8 x i32> %xor.i.i to <8 x float>
+  %2 = bitcast <16 x half> %rhs.coerce to <8 x float>
+  %3 = tail call fast <8 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.256(<8 x float> %1, <8 x float> %2, <8 x float> zeroinitializer, i8 -1) #2
+  %4 = bitcast <8 x float> %3 to <16 x half>
+  ret <16 x half> %4
+}
+
+define dso_local <8 x half> @test3(<8 x half> %lhs.coerce.conj, <8 x half> %rhs.coerce) local_unnamed_addr #0 {
+; CHECK-LABEL: test3:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vfcmulcph %xmm0, %xmm1, %xmm2
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+entry:
+  %0 = bitcast <8 x half> %lhs.coerce.conj to <4 x i32>
+  %xor.i.i = xor <4 x i32> %0, <i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648>
+  %1 = bitcast <4 x i32> %xor.i.i to <4 x float>
+  %2 = bitcast <8 x half> %rhs.coerce to <4 x float>
+  %3 = tail call fast <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float> %1, <4 x float> %2, <4 x float> zeroinitializer, i8 -1) #2
+  %4 = bitcast <4 x float> %3 to <8 x half>
+  ret <8 x half> %4
+}
+
+define dso_local <8 x half> @test4(<8 x half> %lhs.coerce.conj, <8 x half> %rhs.coerce) local_unnamed_addr #0 {
+; CHECK-LABEL: test4:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vfmulcph %xmm0, %xmm1, %xmm2
+; CHECK-NEXT:    vmovaps %xmm2, %xmm0
+; CHECK-NEXT:    retq
+entry:
+  %0 = bitcast <8 x half> %lhs.coerce.conj to <4 x i32>
+  %xor.i.i = xor <4 x i32> %0, <i32 -2147483648, i32 -2147483648, i32 -2147483648, i32 -2147483648>
+  %1 = bitcast <4 x i32> %xor.i.i to <4 x float>
+  %2 = bitcast <8 x half> %rhs.coerce to <4 x float>
+  %3 = tail call fast <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.128(<4 x float> %1, <4 x float> %2, <4 x float> zeroinitializer, i8 -1) #2
+  %4 = bitcast <4 x float> %3 to <8 x half>
+  ret <8 x half> %4
+}
+
+declare <16 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.512(<16 x float>, <16 x float>, <16 x float>, i16, i32 immarg)
+declare <8 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.256(<8 x float>, <8 x float>, <8 x float>, i8)
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float>, <4 x float>, <4 x float>, i8)
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.128(<4 x float>, <4 x float>, <4 x float>, i8)

llvm/test/CodeGen/X86/stack-folding-fp-avx512fp16.ll

define <8 x half> @stack_fold_subsh_int(<8 x half> %a0, <8 x half> %a1) {
   %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
   %2 = extractelement <8 x half> %a0, i32 0
   %3 = extractelement <8 x half> %a1, i32 0
   %4 = fsub half %2, %3
   %5 = insertelement <8 x half> %a0, half %4, i32 0
   ret <8 x half> %5
 }
 
+define <16 x float> @stack_fold_fmulcph(<16 x float> %a0, <16 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_fmulcph:
+  ;CHECK:       vfmulcph {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <16 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.512(<16 x float> %a0, <16 x float> %a1, <16 x float> undef, i16 -1, i32 4)
+  ret <16 x float> %2
+}
+declare <16 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
+
+define <16 x float> @stack_fold_fmulcph_mask(<16 x float> %a0, <16 x float> %a1, <16 x float>* %passthru, i16 %mask) {
+  ;CHECK-LABEL: stack_fold_fmulcph_mask:
+  ;CHECK:       vfmulcph {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{{%k[0-7]}}} {{.*#+}} 64-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load <16 x float>, <16 x float>* %passthru
+  %3 = call <16 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %2, i16 %mask, i32 4)
+  ret <16 x float> %3
+}
+
+define <16 x float> @stack_fold_fmulcph_maskz(<16 x float> %a0, <16 x float> %a1, i16* %mask) {
+  ;CHECK-LABEL: stack_fold_fmulcph_maskz:
+  ;CHECK:       vfmulcph {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 64-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load i16, i16* %mask
+  %3 = call <16 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.512(<16 x float> %a0, <16 x float> %a1, <16 x float> zeroinitializer, i16 %2, i32 4)
+  ret <16 x float> %3
+}
+
+define <16 x float> @stack_fold_fcmulcph(<16 x float> %a0, <16 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_fcmulcph:
+  ;CHECK:       vfcmulcph {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <16 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.512(<16 x float> %a0, <16 x float> %a1, <16 x float> undef, i16 -1, i32 4)
+  ret <16 x float> %2
+}
+declare <16 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
+
+define <16 x float> @stack_fold_fcmulcph_mask(<16 x float> %a0, <16 x float> %a1, <16 x float>* %passthru, i16 %mask) {
+  ;CHECK-LABEL: stack_fold_fcmulcph_mask:
+  ;CHECK:       vfcmulcph {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{{%k[0-7]}}} {{.*#+}} 64-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load <16 x float>, <16 x float>* %passthru
+  %3 = call <16 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %2, i16 %mask, i32 4)
+  ret <16 x float> %3
+}
+
+define <16 x float> @stack_fold_fcmulcph_maskz(<16 x float> %a0, <16 x float> %a1, i16* %mask) {
+  ;CHECK-LABEL: stack_fold_fcmulcph_maskz:
+  ;CHECK:       vfcmulcph {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 64-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load i16, i16* %mask
+  %3 = call <16 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.512(<16 x float> %a0, <16 x float> %a1, <16 x float> zeroinitializer, i16 %2, i32 4)
+  ret <16 x float> %3
+}
+
+define <16 x float> @stack_fold_fmaddcph(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2) {
+  ;CHECK-LABEL: stack_fold_fmaddcph:
+  ;CHECK:       vfmaddcph {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <16 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16 -1, i32 4)
+  ret <16 x float> %2
+}
+declare <16 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
+
+define <16 x float> @stack_fold_fmaddcph_mask(<16 x float>* %p, <16 x float> %a1, <16 x float> %a2, i16 %mask) {
+  ;CHECK-LABEL: stack_fold_fmaddcph_mask:
+  ;CHECK:       vfmaddcph {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{{%k[0-7]}}} {{.*#+}} 64-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %a0 = load <16 x float>, <16 x float>* %p
+  %2 = call <16 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16 %mask, i32 4)
+  ret <16 x float> %2
+}
+
+define <16 x float> @stack_fold_fmaddcph_maskz(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16* %mask) {
+  ;CHECK-LABEL: stack_fold_fmaddcph_maskz:
+  ;CHECK:       vfmaddcph {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 64-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load i16, i16* %mask
+  %3 = call <16 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.512(<16 x float> zeroinitializer, <16 x float> %a1, <16 x float> %a2, i16 %2, i32 4)
+  ret <16 x float> %3
+}
+declare <16 x float> @llvm.x86.avx512fp16.maskz.vfmadd.cph.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
+
+define <16 x float> @stack_fold_fcmaddcph(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2) {
+  ;CHECK-LABEL: stack_fold_fcmaddcph:
+  ;CHECK:       vfcmaddcph {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{.*#+}} 64-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <16 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16 -1, i32 4)
+  ret <16 x float> %2
+}
+declare <16 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
+
+define <16 x float> @stack_fold_fcmaddcph_mask(<16 x float>* %p, <16 x float> %a1, <16 x float> %a2, i16 %mask) {
+  ;CHECK-LABEL: stack_fold_fcmaddcph_mask:
+  ;CHECK:       vfcmaddcph {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{{%k[0-7]}}} {{.*#+}} 64-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %a0 = load <16 x float>, <16 x float>* %p
+  %2 = call <16 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16 %mask, i32 4)
+  ret <16 x float> %2
+}
+
+define <16 x float> @stack_fold_fcmaddcph_maskz(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16* %mask) {
+  ;CHECK-LABEL: stack_fold_fcmaddcph_maskz:
+  ;CHECK:       vfcmaddcph {{-?[0-9]*}}(%rsp), {{%zmm[0-9][0-9]*}}, {{%zmm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 64-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load i16, i16* %mask
+  %3 = call <16 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.512(<16 x float> zeroinitializer, <16 x float> %a1, <16 x float> %a2, i16 %2, i32 4)
+  ret <16 x float> %3
+}
+declare <16 x float> @llvm.x86.avx512fp16.maskz.vfcmadd.cph.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
+
+define <4 x float> @stack_fold_fmulcsh(<4 x float> %a0, <4 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_fmulcsh:
+  ;CHECK:       vfmulcsh {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <4 x float> @llvm.x86.avx512fp16.mask.vfmul.csh(<4 x float> %a0, <4 x float> %a1, <4 x float> undef, i8 -1, i32 4)
+  ret <4 x float> %2
+}
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfmul.csh(<4 x float>, <4 x float>, <4 x float>, i8, i32)
+
+define <4 x float> @stack_fold_fmulcsh_mask(<4 x float> %a0, <4 x float> %a1, <4 x float>* %passthru, i8 %mask) {
+  ;CHECK-LABEL: stack_fold_fmulcsh_mask:
+  ;CHECK:       vfmulcsh {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{{%k[0-7]}}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load <4 x float>, <4 x float>* %passthru
+  %3 = call <4 x float> @llvm.x86.avx512fp16.mask.vfmul.csh(<4 x float> %a0, <4 x float> %a1, <4 x float> %2, i8 %mask, i32 4)
+  ret <4 x float> %3
+}
+
+define <4 x float> @stack_fold_fmulcsh_maskz(<4 x float> %a0, <4 x float> %a1, i8* %mask) {
+  ;CHECK-LABEL: stack_fold_fmulcsh_maskz:
+  ;CHECK:       vfmulcsh {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load i8, i8* %mask
+  %3 = call <4 x float> @llvm.x86.avx512fp16.mask.vfmul.csh(<4 x float> %a0, <4 x float> %a1, <4 x float> zeroinitializer, i8 %2, i32 4)
+  ret <4 x float> %3
+}
+
+define <4 x float> @stack_fold_fcmulcsh(<4 x float> %a0, <4 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_fcmulcsh:
+  ;CHECK:       vfcmulcsh {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.csh(<4 x float> %a0, <4 x float> %a1, <4 x float> undef, i8 -1, i32 4)
+  ret <4 x float> %2
+}
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.csh(<4 x float>, <4 x float>, <4 x float>, i8, i32)
+
+define <4 x float> @stack_fold_fcmulcsh_mask(<4 x float> %a0, <4 x float> %a1, <4 x float>* %passthru, i8 %mask) {
+  ;CHECK-LABEL: stack_fold_fcmulcsh_mask:
+  ;CHECK:       vfcmulcsh {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{{%k[0-7]}}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load <4 x float>, <4 x float>* %passthru
+  %3 = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.csh(<4 x float> %a0, <4 x float> %a1, <4 x float> %2, i8 %mask, i32 4)
+  ret <4 x float> %3
+}
+
+define <4 x float> @stack_fold_fcmulcsh_maskz(<4 x float> %a0, <4 x float> %a1, i8* %mask) {
+  ;CHECK-LABEL: stack_fold_fcmulcsh_maskz:
+  ;CHECK:       vfcmulcsh {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load i8, i8* %mask
+  %3 = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.csh(<4 x float> %a0, <4 x float> %a1, <4 x float> zeroinitializer, i8 %2, i32 4)
+  ret <4 x float> %3
+}
+
+define <4 x float> @stack_fold_fmaddcsh(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2) {
+  ;CHECK-LABEL: stack_fold_fmaddcsh:
+  ;CHECK:       vfmaddcsh {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.csh(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 -1, i32 4)
+  ret <4 x float> %2
+}
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.csh(<4 x float>, <4 x float>, <4 x float>, i8, i32)
+
+define <4 x float> @stack_fold_fmaddcsh_mask(<4 x float>* %p, <4 x float> %a1, <4 x float> %a2, i8 %mask) {
+  ;CHECK-LABEL: stack_fold_fmaddcsh_mask:
+  ;CHECK:       vfmaddcsh {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{{%k[0-7]}}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %a0 = load <4 x float>, <4 x float>* %p
+  %2 = call <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.csh(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask, i32 4)
+  ret <4 x float> %2
+}
+
+define <4 x float> @stack_fold_fmaddcsh_maskz(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8* %mask) {
+  ;CHECK-LABEL: stack_fold_fmaddcsh_maskz:
+  ;CHECK:       vfmaddcsh {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load i8, i8* %mask
+  %3 = call <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.csh(<4 x float> zeroinitializer, <4 x float> %a1, <4 x float> %a2, i8 %2, i32 4)
+  ret <4 x float> %3
+}
+declare <4 x float> @llvm.x86.avx512fp16.maskz.vfmadd.csh(<4 x float>, <4 x float>, <4 x float>, i8, i32)
+
+define <4 x float> @stack_fold_fcmaddcsh(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2) {
+  ;CHECK-LABEL: stack_fold_fcmaddcsh:
+  ;CHECK:       vfcmaddcsh {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.csh(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 -1, i32 4)
+  ret <4 x float> %2
+}
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.csh(<4 x float>, <4 x float>, <4 x float>, i8, i32)
+
+define <4 x float> @stack_fold_fcmaddcsh_mask(<4 x float>* %p, <4 x float> %a1, <4 x float> %a2, i8 %mask) {
+  ;CHECK-LABEL: stack_fold_fcmaddcsh_mask:
+  ;CHECK:       vfcmaddcsh {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{{%k[0-7]}}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %a0 = load <4 x float>, <4 x float>* %p
+  %2 = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.csh(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask, i32 4)
+  ret <4 x float> %2
+}
+
+define <4 x float> @stack_fold_fcmaddcsh_maskz(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8* %mask) {
+  ;CHECK-LABEL: stack_fold_fcmaddcsh_maskz:
+  ;CHECK:       vfcmaddcsh {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load i8, i8* %mask
+  %3 = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.csh(<4 x float> zeroinitializer, <4 x float> %a1, <4 x float> %a2, i8 %2, i32 4)
+  ret <4 x float> %3
+}
+declare <4 x float> @llvm.x86.avx512fp16.maskz.vfcmadd.csh(<4 x float>, <4 x float>, <4 x float>, i8, i32)
+
 attributes #0 = { "unsafe-fp-math"="false" }
 attributes #1 = { "unsafe-fp-math"="true" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" }

llvm/test/CodeGen/X86/stack-folding-fp-avx512fp16vl.ll

 define <16 x half> @stack_fold_subph_ymm(<16 x half> %a0, <16 x half> %a1) {
   ;CHECK-LABEL: stack_fold_subph_ymm
   ;CHECK:       vsubph {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
   %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
   %2 = fsub <16 x half> %a0, %a1
   ret <16 x half> %2
 }
 
+define <4 x float> @stack_fold_fmulc(<4 x float> %a0, <4 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_fmulc:
+  ;CHECK:       vfmulcph {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float> %a0, <4 x float> %a1, <4 x float> undef, i8 -1)
+  ret <4 x float> %2
+}
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float>, <4 x float>, <4 x float>, i8)
+
+define <4 x float> @stack_fold_fmulc_mask(<4 x float> %a0, <4 x float> %a1, <4 x float>* %passthru, i8 %mask) {
+  ;CHECK-LABEL: stack_fold_fmulc_mask:
+  ;CHECK:       vfmulcph {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{{%k[0-7]}}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load <4 x float>, <4 x float>* %passthru
+  %3 = call <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %2, i8 %mask)
+  ret <4 x float> %3
+}
+
+define <4 x float> @stack_fold_fmulc_maskz(<4 x float> %a0, <4 x float> %a1, i8* %mask) {
+  ;CHECK-LABEL: stack_fold_fmulc_maskz:
+  ;CHECK:       vfmulcph {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load i8, i8* %mask
+  %3 = call <4 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.128(<4 x float> %a0, <4 x float> %a1, <4 x float> zeroinitializer, i8 %2)
+  ret <4 x float> %3
+}
+
+define <4 x float> @stack_fold_fcmulc(<4 x float> %a0, <4 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_fcmulc:
+  ;CHECK:       vfcmulcph {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.128(<4 x float> %a0, <4 x float> %a1, <4 x float> undef, i8 -1)
+  ret <4 x float> %2
+}
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.128(<4 x float>, <4 x float>, <4 x float>, i8)
+
+define <4 x float> @stack_fold_fcmulc_mask(<4 x float> %a0, <4 x float> %a1, <4 x float>* %passthru, i8 %mask) {
+  ;CHECK-LABEL: stack_fold_fcmulc_mask:
+  ;CHECK:       vfcmulcph {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{{%k[0-7]}}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load <4 x float>, <4 x float>* %passthru
+  %3 = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %2, i8 %mask)
+  ret <4 x float> %3
+}
+
+define <4 x float> @stack_fold_fcmulc_maskz(<4 x float> %a0, <4 x float> %a1, i8* %mask) {
+  ;CHECK-LABEL: stack_fold_fcmulc_maskz:
+  ;CHECK:       vfcmulcph {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load i8, i8* %mask
+  %3 = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.128(<4 x float> %a0, <4 x float> %a1, <4 x float> zeroinitializer, i8 %2)
+  ret <4 x float> %3
+}
+
+define <4 x float> @stack_fold_fmaddc(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2) {
+  ;CHECK-LABEL: stack_fold_fmaddc:
+  ;CHECK:       vfmaddcph {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 -1)
+  ret <4 x float> %2
+}
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.128(<4 x float>, <4 x float>, <4 x float>, i8)
+
+define <4 x float> @stack_fold_fmaddc_mask(<4 x float>* %p, <4 x float> %a1, <4 x float> %a2, i8 %mask) {
+  ;CHECK-LABEL: stack_fold_fmaddc_mask:
+  ;CHECK:       vfmaddcph {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{{%k[0-7]}}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %a0 = load <4 x float>, <4 x float>* %p
+  %2 = call <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask)
+  ret <4 x float> %2
+}
+
+define <4 x float> @stack_fold_fmaddc_maskz(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8* %mask) {
+  ;CHECK-LABEL: stack_fold_fmaddc_maskz:
+  ;CHECK:       vfmaddcph {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load i8, i8* %mask
+  %3 = call <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.128(<4 x float> zeroinitializer, <4 x float> %a1, <4 x float> %a2, i8 %2)
+  ret <4 x float> %3
+}
+declare <4 x float> @llvm.x86.avx512fp16.maskz.vfmadd.cph.128(<4 x float>, <4 x float>, <4 x float>, i8)
+
+define <4 x float> @stack_fold_fcmaddc(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2) {
+  ;CHECK-LABEL: stack_fold_fcmaddc:
+  ;CHECK:       vfcmaddcph {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 -1)
+  ret <4 x float> %2
+}
+declare <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.128(<4 x float>, <4 x float>, <4 x float>, i8)
+
+define <4 x float> @stack_fold_fcmaddc_mask(<4 x float>* %p, <4 x float> %a1, <4 x float> %a2, i8 %mask) {
+  ;CHECK-LABEL: stack_fold_fcmaddc_mask:
+  ;CHECK:       vfcmaddcph {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{{%k[0-7]}}} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %a0 = load <4 x float>, <4 x float>* %p
+  %2 = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask)
+  ret <4 x float> %2
+}
+
+define <4 x float> @stack_fold_fcmaddc_maskz(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8* %mask) {
+  ;CHECK-LABEL: stack_fold_fcmaddc_maskz:
+  ;CHECK:       vfcmaddcph {{-?[0-9]*}}(%rsp), {{%xmm[0-9][0-9]*}}, {{%xmm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 16-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load i8, i8* %mask
+  %3 = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.128(<4 x float> zeroinitializer, <4 x float> %a1, <4 x float> %a2, i8 %2)
+  ret <4 x float> %3
+}
+declare <4 x float> @llvm.x86.avx512fp16.maskz.vfcmadd.cph.128(<4 x float>, <4 x float>, <4 x float>, i8)
+
+define <8 x float> @stack_fold_fmulc_ymm(<8 x float> %a0, <8 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_fmulc_ymm:
+  ;CHECK:       vfmulcph {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <8 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.256(<8 x float> %a0, <8 x float> %a1, <8 x float> undef, i8 -1)
+  ret <8 x float> %2
+}
+declare <8 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.256(<8 x float>, <8 x float>, <8 x float>, i8)
+
+define <8 x float> @stack_fold_fmulc_mask_ymm(<8 x float> %a0, <8 x float> %a1, <8 x float>* %passthru, i8 %mask) {
+  ;CHECK-LABEL: stack_fold_fmulc_mask_ymm:
+  ;CHECK:       vfmulcph {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{{%k[0-7]}}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load <8 x float>, <8 x float>* %passthru
+  %3 = call <8 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %2, i8 %mask)
+  ret <8 x float> %3
+}
+
+define <8 x float> @stack_fold_fmulc_maskz_ymm(<8 x float> %a0, <8 x float> %a1, i8* %mask) {
+  ;CHECK-LABEL: stack_fold_fmulc_maskz_ymm:
+  ;CHECK:       vfmulcph {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load i8, i8* %mask
+  %3 = call <8 x float> @llvm.x86.avx512fp16.mask.vfmul.cph.256(<8 x float> %a0, <8 x float> %a1, <8 x float> zeroinitializer, i8 %2)
+  ret <8 x float> %3
+}
+
+define <8 x float> @stack_fold_fcmulc_ymm(<8 x float> %a0, <8 x float> %a1) {
+  ;CHECK-LABEL: stack_fold_fcmulc_ymm:
+  ;CHECK:       vfcmulcph {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <8 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.256(<8 x float> %a0, <8 x float> %a1, <8 x float> undef, i8 -1)
+  ret <8 x float> %2
+}
+declare <8 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.256(<8 x float>, <8 x float>, <8 x float>, i8)
+
+define <8 x float> @stack_fold_fcmulc_mask_ymm(<8 x float> %a0, <8 x float> %a1, <8 x float>* %passthru, i8 %mask) {
+  ;CHECK-LABEL: stack_fold_fcmulc_mask_ymm:
+  ;CHECK:       vfcmulcph {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{{%k[0-7]}}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load <8 x float>, <8 x float>* %passthru
+  %3 = call <8 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %2, i8 %mask)
+  ret <8 x float> %3
+}
+
+define <8 x float> @stack_fold_fcmulc_maskz_ymm(<8 x float> %a0, <8 x float> %a1, i8* %mask) {
+  ;CHECK-LABEL: stack_fold_fcmulc_maskz_ymm:
+  ;CHECK:       vfcmulcph {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load i8, i8* %mask
+  %3 = call <8 x float> @llvm.x86.avx512fp16.mask.vfcmul.cph.256(<8 x float> %a0, <8 x float> %a1, <8 x float> zeroinitializer, i8 %2)
+  ret <8 x float> %3
+}
+
+define <8 x float> @stack_fold_fmaddc_ymm(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2) {
+  ;CHECK-LABEL: stack_fold_fmaddc_ymm:
+  ;CHECK:       vfmaddcph {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <8 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2, i8 -1)
+  ret <8 x float> %2
+}
+declare <8 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.256(<8 x float>, <8 x float>, <8 x float>, i8)
+
+define <8 x float> @stack_fold_fmaddc_mask_ymm(<8 x float>* %p, <8 x float> %a1, <8 x float> %a2, i8 %mask) {
+  ;CHECK-LABEL: stack_fold_fmaddc_mask_ymm:
+  ;CHECK:       vfmaddcph {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{{%k[0-7]}}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %a0 = load <8 x float>, <8 x float>* %p
+  %2 = call <8 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2, i8 %mask)
+  ret <8 x float> %2
+}
+
+define <8 x float> @stack_fold_fmaddc_maskz_ymm(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2, i8* %mask) {
+  ;CHECK-LABEL: stack_fold_fmaddc_maskz_ymm:
+  ;CHECK:       vfmaddcph {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load i8, i8* %mask
+  %3 = call <8 x float> @llvm.x86.avx512fp16.mask.vfmadd.cph.256(<8 x float> zeroinitializer, <8 x float> %a1, <8 x float> %a2, i8 %2)
+  ret <8 x float> %3
+}
+declare <8 x float> @llvm.x86.avx512fp16.maskz.vfmadd.cph.256(<8 x float>, <8 x float>, <8 x float>, i8)
+
+define <8 x float> @stack_fold_fcmaddc_ymm(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2) {
+  ;CHECK-LABEL: stack_fold_fcmaddc_ymm:
+  ;CHECK:       vfcmaddcph {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = call <8 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2, i8 -1)
+  ret <8 x float> %2
+}
+declare <8 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.256(<8 x float>, <8 x float>, <8 x float>, i8)
+
+define <8 x float> @stack_fold_fcmaddc_mask_ymm(<8 x float>* %p, <8 x float> %a1, <8 x float> %a2, i8 %mask) {
+  ;CHECK-LABEL: stack_fold_fcmaddc_mask_ymm:
+  ;CHECK:       vfcmaddcph {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{{%k[0-7]}}} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %a0 = load <8 x float>, <8 x float>* %p
+  %2 = call <8 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2, i8 %mask)
+  ret <8 x float> %2
+}
+
+define <8 x float> @stack_fold_fcmaddc_maskz_ymm(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2, i8* %mask) {
+  ;CHECK-LABEL: stack_fold_fcmaddc_maskz_ymm:
+  ;CHECK:       vfcmaddcph {{-?[0-9]*}}(%rsp), {{%ymm[0-9][0-9]*}}, {{%ymm[0-9][0-9]*}} {{{%k[0-7]}}} {z} {{.*#+}} 32-byte Folded Reload
+  %1 = tail call <2 x i64> asm sideeffect "nop", "=x,~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{flags}"()
+  %2 = load i8, i8* %mask
+  %3 = call <8 x float> @llvm.x86.avx512fp16.mask.vfcmadd.cph.256(<8 x float> zeroinitializer, <8 x float> %a1, <8 x float> %a2, i8 %2)
+  ret <8 x float> %3
+}
+declare <8 x float> @llvm.x86.avx512fp16.maskz.vfcmadd.cph.256(<8 x float>, <8 x float>, <8 x float>, i8)
+
 attributes #0 = { "unsafe-fp-math"="false" }
 attributes #1 = { "unsafe-fp-math"="true" }

llvm/test/MC/Disassembler/X86/avx512fp16.txt

 
 # ATT:   vfnmsub231sh  254(%rcx), %xmm29, %xmm30
 # INTEL: vfnmsub231sh xmm30, xmm29, word ptr [rcx + 254]
 0x62,0x66,0x15,0x00,0xbf,0x71,0x7f
 
 # ATT:   vfnmsub231sh  -256(%rdx), %xmm29, %xmm30 {%k7} {z}
 # INTEL: vfnmsub231sh xmm30 {k7} {z}, xmm29, word ptr [rdx - 256]
 0x62,0x66,0x15,0x87,0xbf,0x72,0x80
+
+# ATT:   vfcmaddcph %zmm28, %zmm29, %zmm30
+# INTEL: vfcmaddcph zmm30, zmm29, zmm28
+0x62,0x06,0x17,0x40,0x56,0xf4
+
+# ATT:   vfcmaddcph {rn-sae}, %zmm28, %zmm29, %zmm30
+# INTEL: vfcmaddcph zmm30, zmm29, zmm28, {rn-sae}
+0x62,0x06,0x17,0x10,0x56,0xf4
+
+# ATT:   vfcmaddcph  268435456(%rbp,%r14,8), %zmm29, %zmm30 {%k7}
+# INTEL: vfcmaddcph zmm30 {k7}, zmm29, zmmword ptr [rbp + 8*r14 + 268435456]
+0x62,0x26,0x17,0x47,0x56,0xb4,0xf5,0x00,0x00,0x00,0x10
+
+# ATT:   vfcmaddcph  (%r9){1to16}, %zmm29, %zmm30
+# INTEL: vfcmaddcph zmm30, zmm29, dword ptr [r9]{1to16}
+0x62,0x46,0x17,0x50,0x56,0x31
+
+# ATT:   vfcmaddcph  8128(%rcx), %zmm29, %zmm30
+# INTEL: vfcmaddcph zmm30, zmm29, zmmword ptr [rcx + 8128]
+0x62,0x66,0x17,0x40,0x56,0x71,0x7f
+
+# ATT:   vfcmaddcph  -512(%rdx){1to16}, %zmm29, %zmm30 {%k7} {z}
+# INTEL: vfcmaddcph zmm30 {k7} {z}, zmm29, dword ptr [rdx - 512]{1to16}
+0x62,0x66,0x17,0xd7,0x56,0x72,0x80
+
+# ATT:   vfcmaddcsh %xmm28, %xmm29, %xmm30
+# INTEL: vfcmaddcsh xmm30, xmm29, xmm28
+0x62,0x06,0x17,0x00,0x57,0xf4
+
+# ATT:   vfcmaddcsh {rn-sae}, %xmm28, %xmm29, %xmm30
+# INTEL: vfcmaddcsh xmm30, xmm29, xmm28, {rn-sae}
+0x62,0x06,0x17,0x10,0x57,0xf4
+
+# ATT:   vfcmaddcsh  268435456(%rbp,%r14,8), %xmm29, %xmm30 {%k7}
+# INTEL: vfcmaddcsh xmm30 {k7}, xmm29, dword ptr [rbp + 8*r14 + 268435456]
+0x62,0x26,0x17,0x07,0x57,0xb4,0xf5,0x00,0x00,0x00,0x10
+
+# ATT:   vfcmaddcsh  (%r9), %xmm29, %xmm30
+# INTEL: vfcmaddcsh xmm30, xmm29, dword ptr [r9]
+0x62,0x46,0x17,0x00,0x57,0x31
+
+# ATT:   vfcmaddcsh  508(%rcx), %xmm29, %xmm30
+# INTEL: vfcmaddcsh xmm30, xmm29, dword ptr [rcx + 508]
+0x62,0x66,0x17,0x00,0x57,0x71,0x7f
+
+# ATT:   vfcmaddcsh  -512(%rdx), %xmm29, %xmm30 {%k7} {z}
+# INTEL: vfcmaddcsh xmm30 {k7} {z}, xmm29, dword ptr [rdx - 512]
+0x62,0x66,0x17,0x87,0x57,0x72,0x80
+
+# ATT:   vfcmulcph %zmm28, %zmm29, %zmm30
+# INTEL: vfcmulcph zmm30, zmm29, zmm28
+0x62,0x06,0x17,0x40,0xd6,0xf4
+
+# ATT:   vfcmulcph {rn-sae}, %zmm28, %zmm29, %zmm30
+# INTEL: vfcmulcph zmm30, zmm29, zmm28, {rn-sae}
+0x62,0x06,0x17,0x10,0xd6,0xf4
+
+# ATT:   vfcmulcph  268435456(%rbp,%r14,8), %zmm29, %zmm30 {%k7}
+# INTEL: vfcmulcph zmm30 {k7}, zmm29, zmmword ptr [rbp + 8*r14 + 268435456]
+0x62,0x26,0x17,0x47,0xd6,0xb4,0xf5,0x00,0x00,0x00,0x10
+
+# ATT:   vfcmulcph  (%r9){1to16}, %zmm29, %zmm30
+# INTEL: vfcmulcph zmm30, zmm29, dword ptr [r9]{1to16}
+0x62,0x46,0x17,0x50,0xd6,0x31
+
+# ATT:   vfcmulcph  8128(%rcx), %zmm29, %zmm30
+# INTEL: vfcmulcph zmm30, zmm29, zmmword ptr [rcx + 8128]
+0x62,0x66,0x17,0x40,0xd6,0x71,0x7f
+
+# ATT:   vfcmulcph  -512(%rdx){1to16}, %zmm29, %zmm30 {%k7} {z}
+# INTEL: vfcmulcph zmm30 {k7} {z}, zmm29, dword ptr [rdx - 512]{1to16}
+0x62,0x66,0x17,0xd7,0xd6,0x72,0x80
+
+# ATT:   vfcmulcsh %xmm28, %xmm29, %xmm30
+# INTEL: vfcmulcsh xmm30, xmm29, xmm28
+0x62,0x06,0x17,0x00,0xd7,0xf4
+
+# ATT:   vfcmulcsh {rn-sae}, %xmm28, %xmm29, %xmm30
+# INTEL: vfcmulcsh xmm30, xmm29, xmm28, {rn-sae}
+0x62,0x06,0x17,0x10,0xd7,0xf4
+
+# ATT:   vfcmulcsh  268435456(%rbp,%r14,8), %xmm29, %xmm30 {%k7}
+# INTEL: vfcmulcsh xmm30 {k7}, xmm29, dword ptr [rbp + 8*r14 + 268435456]
+0x62,0x26,0x17,0x07,0xd7,0xb4,0xf5,0x00,0x00,0x00,0x10
+
+# ATT:   vfcmulcsh  (%r9), %xmm29, %xmm30
+# INTEL: vfcmulcsh xmm30, xmm29, dword ptr [r9]
+0x62,0x46,0x17,0x00,0xd7,0x31
+
+# ATT:   vfcmulcsh  508(%rcx), %xmm29, %xmm30
+# INTEL: vfcmulcsh xmm30, xmm29, dword ptr [rcx + 508]
+0x62,0x66,0x17,0x00,0xd7,0x71,0x7f
+
+# ATT:   vfcmulcsh  -512(%rdx), %xmm29, %xmm30 {%k7} {z}
+# INTEL: vfcmulcsh xmm30 {k7} {z}, xmm29, dword ptr [rdx - 512]
+0x62,0x66,0x17,0x87,0xd7,0x72,0x80
+
+# ATT:   vfmaddcph %zmm28, %zmm29, %zmm30
+# INTEL: vfmaddcph zmm30, zmm29, zmm28
+0x62,0x06,0x16,0x40,0x56,0xf4
+
+# ATT:   vfmaddcph {rn-sae}, %zmm28, %zmm29, %zmm30
+# INTEL: vfmaddcph zmm30, zmm29, zmm28, {rn-sae}
+0x62,0x06,0x16,0x10,0x56,0xf4
+
+# ATT:   vfmaddcph  268435456(%rbp,%r14,8), %zmm29, %zmm30 {%k7}
+# INTEL: vfmaddcph zmm30 {k7}, zmm29, zmmword ptr [rbp + 8*r14 + 268435456]
+0x62,0x26,0x16,0x47,0x56,0xb4,0xf5,0x00,0x00,0x00,0x10
+
+# ATT:   vfmaddcph  (%r9){1to16}, %zmm29, %zmm30
+# INTEL: vfmaddcph zmm30, zmm29, dword ptr [r9]{1to16}
+0x62,0x46,0x16,0x50,0x56,0x31
+
+# ATT:   vfmaddcph  8128(%rcx), %zmm29, %zmm30
+# INTEL: vfmaddcph zmm30, zmm29, zmmword ptr [rcx + 8128]
+0x62,0x66,0x16,0x40,0x56,0x71,0x7f
+
+# ATT:   vfmaddcph  -512(%rdx){1to16}, %zmm29, %zmm30 {%k7} {z}
+# INTEL: vfmaddcph zmm30 {k7} {z}, zmm29, dword ptr [rdx - 512]{1to16}
+0x62,0x66,0x16,0xd7,0x56,0x72,0x80
+
+# ATT:   vfmaddcsh %xmm28, %xmm29, %xmm30
+# INTEL: vfmaddcsh xmm30, xmm29, xmm28
+0x62,0x06,0x16,0x00,0x57,0xf4
+
+# ATT:   vfmaddcsh {rn-sae}, %xmm28, %xmm29, %xmm30
+# INTEL: vfmaddcsh xmm30, xmm29, xmm28, {rn-sae}
+0x62,0x06,0x16,0x10,0x57,0xf4
+
+# ATT:   vfmaddcsh  268435456(%rbp,%r14,8), %xmm29, %xmm30 {%k7}
+# INTEL: vfmaddcsh xmm30 {k7}, xmm29, dword ptr [rbp + 8*r14 + 268435456]
+0x62,0x26,0x16,0x07,0x57,0xb4,0xf5,0x00,0x00,0x00,0x10
+
+# ATT:   vfmaddcsh  (%r9), %xmm29, %xmm30
+# INTEL: vfmaddcsh xmm30, xmm29, dword ptr [r9]
+0x62,0x46,0x16,0x00,0x57,0x31
+
+# ATT:   vfmaddcsh  508(%rcx), %xmm29, %xmm30
+# INTEL: vfmaddcsh xmm30, xmm29, dword ptr [rcx + 508]
+0x62,0x66,0x16,0x00,0x57,0x71,0x7f
+
+# ATT:   vfmaddcsh  -512(%rdx), %xmm29, %xmm30 {%k7} {z}
+# INTEL: vfmaddcsh xmm30 {k7} {z}, xmm29, dword ptr [rdx - 512]
+0x62,0x66,0x16,0x87,0x57,0x72,0x80
+
+# ATT:   vfmulcph %zmm28, %zmm29, %zmm30
+# INTEL: vfmulcph zmm30, zmm29, zmm28
+0x62,0x06,0x16,0x40,0xd6,0xf4
+
+# ATT:   vfmulcph {rn-sae}, %zmm28, %zmm29, %zmm30
+# INTEL: vfmulcph zmm30, zmm29, zmm28, {rn-sae}
+0x62,0x06,0x16,0x10,0xd6,0xf4
+
+# ATT:   vfmulcph  268435456(%rbp,%r14,8), %zmm29, %zmm30 {%k7}
+# INTEL: vfmulcph zmm30 {k7}, zmm29, zmmword ptr [rbp + 8*r14 + 268435456]
+0x62,0x26,0x16,0x47,0xd6,0xb4,0xf5,0x00,0x00,0x00,0x10
+
+# ATT:   vfmulcph  (%r9){1to16}, %zmm29, %zmm30
+# INTEL: vfmulcph zmm30, zmm29, dword ptr [r9]{1to16}
+0x62,0x46,0x16,0x50,0xd6,0x31
+
+# ATT:   vfmulcph  8128(%rcx), %zmm29, %zmm30
+# INTEL: vfmulcph zmm30, zmm29, zmmword ptr [rcx + 8128]
+0x62,0x66,0x16,0x40,0xd6,0x71,0x7f
+
+# ATT:   vfmulcph  -512(%rdx){1to16}, %zmm29, %zmm30 {%k7} {z}
+# INTEL: vfmulcph zmm30 {k7} {z}, zmm29, dword ptr [rdx - 512]{1to16}
+0x62,0x66,0x16,0xd7,0xd6,0x72,0x80
+
+# ATT:   vfmulcsh %xmm28, %xmm29, %xmm30
+# INTEL: vfmulcsh xmm30, xmm29, xmm28
+0x62,0x06,0x16,0x00,0xd7,0xf4
+
+# ATT:   vfmulcsh {rn-sae}, %xmm28, %xmm29, %xmm30
+# INTEL: vfmulcsh xmm30, xmm29, xmm28, {rn-sae}
+0x62,0x06,0x16,0x10,0xd7,0xf4
+
+# ATT:   vfmulcsh  268435456(%rbp,%r14,8), %xmm29, %xmm30 {%k7}
+# INTEL: vfmulcsh xmm30 {k7}, xmm29, dword ptr [rbp + 8*r14 + 268435456]
+0x62,0x26,0x16,0x07,0xd7,0xb4,0xf5,0x00,0x00,0x00,0x10
+
+# ATT:   vfmulcsh  (%r9), %xmm29, %xmm30
+# INTEL: vfmulcsh xmm30, xmm29, dword ptr [r9]
+0x62,0x46,0x16,0x00,0xd7,0x31
+
+# ATT:   vfmulcsh  508(%rcx), %xmm29, %xmm30
+# INTEL: vfmulcsh xmm30, xmm29, dword ptr [rcx + 508]
+0x62,0x66,0x16,0x00,0xd7,0x71,0x7f
+
+# ATT:   vfmulcsh  -512(%rdx), %xmm29, %xmm30 {%k7} {z}
+# INTEL: vfmulcsh xmm30 {k7} {z}, xmm29, dword ptr [rdx - 512]
+0x62,0x66,0x16,0x87,0xd7,0x72,0x80

llvm/test/MC/Disassembler/X86/avx512fp16vl.txt

 
 # ATT:   vfnmsub231ph  2032(%ecx), %xmm5, %xmm6
 # INTEL: vfnmsub231ph xmm6, xmm5, xmmword ptr [ecx + 2032]
 0x62,0xf6,0x55,0x08,0xbe,0x71,0x7f
 
 # ATT:   vfnmsub231ph  -256(%edx){1to8}, %xmm5, %xmm6 {%k7} {z}
 # INTEL: vfnmsub231ph xmm6 {k7} {z}, xmm5, word ptr [edx - 256]{1to8}
 0x62,0xf6,0x55,0x9f,0xbe,0x72,0x80
+
+# ATT:   vfcmaddcph %ymm4, %ymm5, %ymm6
+# INTEL: vfcmaddcph ymm6, ymm5, ymm4
+0x62,0xf6,0x57,0x28,0x56,0xf4
+
+# ATT:   vfcmaddcph %xmm4, %xmm5, %xmm6
+# INTEL: vfcmaddcph xmm6, xmm5, xmm4
+0x62,0xf6,0x57,0x08,0x56,0xf4
+
+# ATT:   vfcmaddcph  268435456(%esp,%esi,8), %ymm5, %ymm6 {%k7}
+# INTEL: vfcmaddcph ymm6 {k7}, ymm5, ymmword ptr [esp + 8*esi + 268435456]
+0x62,0xf6,0x57,0x2f,0x56,0xb4,0xf4,0x00,0x00,0x00,0x10
+
+# ATT:   vfcmaddcph  (%ecx){1to8}, %ymm5, %ymm6
+# INTEL: vfcmaddcph ymm6, ymm5, dword ptr [ecx]{1to8}
+0x62,0xf6,0x57,0x38,0x56,0x31
+
+# ATT:   vfcmaddcph  4064(%ecx), %ymm5, %ymm6
+# INTEL: vfcmaddcph ymm6, ymm5, ymmword ptr [ecx + 4064]
+0x62,0xf6,0x57,0x28,0x56,0x71,0x7f
+
+# ATT:   vfcmaddcph  -512(%edx){1to8}, %ymm5, %ymm6 {%k7} {z}
+# INTEL: vfcmaddcph ymm6 {k7} {z}, ymm5, dword ptr [edx - 512]{1to8}
+0x62,0xf6,0x57,0xbf,0x56,0x72,0x80
+
+# ATT:   vfcmaddcph  268435456(%esp,%esi,8), %xmm5, %xmm6 {%k7}
+# INTEL: vfcmaddcph xmm6 {k7}, xmm5, xmmword ptr [esp + 8*esi + 268435456]
+0x62,0xf6,0x57,0x0f,0x56,0xb4,0xf4,0x00,0x00,0x00,0x10
+
+# ATT:   vfcmaddcph  (%ecx){1to4}, %xmm5, %xmm6
+# INTEL: vfcmaddcph xmm6, xmm5, dword ptr [ecx]{1to4}
+0x62,0xf6,0x57,0x18,0x56,0x31
+
+# ATT:   vfcmaddcph  2032(%ecx), %xmm5, %xmm6
+# INTEL: vfcmaddcph xmm6, xmm5, xmmword ptr [ecx + 2032]
+0x62,0xf6,0x57,0x08,0x56,0x71,0x7f
+
+# ATT:   vfcmaddcph  -512(%edx){1to4}, %xmm5, %xmm6 {%k7} {z}
+# INTEL: vfcmaddcph xmm6 {k7} {z}, xmm5, dword ptr [edx - 512]{1to4}
+0x62,0xf6,0x57,0x9f,0x56,0x72,0x80
+
+# ATT:   vfcmulcph %ymm4, %ymm5, %ymm6
+# INTEL: vfcmulcph ymm6, ymm5, ymm4
+0x62,0xf6,0x57,0x28,0xd6,0xf4
+
+# ATT:   vfcmulcph %xmm4, %xmm5, %xmm6
+# INTEL: vfcmulcph xmm6, xmm5, xmm4
+0x62,0xf6,0x57,0x08,0xd6,0xf4
+
+# ATT:   vfcmulcph  268435456(%esp,%esi,8), %ymm5, %ymm6 {%k7}
+# INTEL: vfcmulcph ymm6 {k7}, ymm5, ymmword ptr [esp + 8*esi + 268435456]
+0x62,0xf6,0x57,0x2f,0xd6,0xb4,0xf4,0x00,0x00,0x00,0x10
+
+# ATT:   vfcmulcph  (%ecx){1to8}, %ymm5, %ymm6
+# INTEL: vfcmulcph ymm6, ymm5, dword ptr [ecx]{1to8}
+0x62,0xf6,0x57,0x38,0xd6,0x31
+
+# ATT:   vfcmulcph  4064(%ecx), %ymm5, %ymm6
+# INTEL: vfcmulcph ymm6, ymm5, ymmword ptr [ecx + 4064]
+0x62,0xf6,0x57,0x28,0xd6,0x71,0x7f
+
+# ATT:   vfcmulcph  -512(%edx){1to8}, %ymm5, %ymm6 {%k7} {z}
+# INTEL: vfcmulcph ymm6 {k7} {z}, ymm5, dword ptr [edx - 512]{1to8}
+0x62,0xf6,0x57,0xbf,0xd6,0x72,0x80
+
+# ATT:   vfcmulcph  268435456(%esp,%esi,8), %xmm5, %xmm6 {%k7}
+# INTEL: vfcmulcph xmm6 {k7}, xmm5, xmmword ptr [esp + 8*esi + 268435456]
+0x62,0xf6,0x57,0x0f,0xd6,0xb4,0xf4,0x00,0x00,0x00,0x10
+
+# ATT:   vfcmulcph  (%ecx){1to4}, %xmm5, %xmm6
+# INTEL: vfcmulcph xmm6, xmm5, dword ptr [ecx]{1to4}
+0x62,0xf6,0x57,0x18,0xd6,0x31
+
+# ATT:   vfcmulcph  2032(%ecx), %xmm5, %xmm6
+# INTEL: vfcmulcph xmm6, xmm5, xmmword ptr [ecx + 2032]
+0x62,0xf6,0x57,0x08,0xd6,0x71,0x7f
+
+# ATT:   vfcmulcph  -512(%edx){1to4}, %xmm5, %xmm6 {%k7} {z}
+# INTEL: vfcmulcph xmm6 {k7} {z}, xmm5, dword ptr [edx - 512]{1to4}
+0x62,0xf6,0x57,0x9f,0xd6,0x72,0x80
+
+# ATT:   vfmaddcph %ymm4, %ymm5, %ymm6
+# INTEL: vfmaddcph ymm6, ymm5, ymm4
+0x62,0xf6,0x56,0x28,0x56,0xf4
+
+# ATT:   vfmaddcph %xmm4, %xmm5, %xmm6
+# INTEL: vfmaddcph xmm6, xmm5, xmm4
+0x62,0xf6,0x56,0x08,0x56,0xf4
+
+# ATT:   vfmaddcph  268435456(%esp,%esi,8), %ymm5, %ymm6 {%k7}
+# INTEL: vfmaddcph ymm6 {k7}, ymm5, ymmword ptr [esp + 8*esi + 268435456]
+0x62,0xf6,0x56,0x2f,0x56,0xb4,0xf4,0x00,0x00,0x00,0x10
+
+# ATT:   vfmaddcph  (%ecx){1to8}, %ymm5, %ymm6
+# INTEL: vfmaddcph ymm6, ymm5, dword ptr [ecx]{1to8}
+0x62,0xf6,0x56,0x38,0x56,0x31
+
+# ATT:   vfmaddcph  4064(%ecx), %ymm5, %ymm6
+# INTEL: vfmaddcph ymm6, ymm5, ymmword ptr [ecx + 4064]
+0x62,0xf6,0x56,0x28,0x56,0x71,0x7f
+
+# ATT:   vfmaddcph  -512(%edx){1to8}, %ymm5, %ymm6 {%k7} {z}
+# INTEL: vfmaddcph ymm6 {k7} {z}, ymm5, dword ptr [edx - 512]{1to8}
+0x62,0xf6,0x56,0xbf,0x56,0x72,0x80
+
+# ATT:   vfmaddcph  268435456(%esp,%esi,8), %xmm5, %xmm6 {%k7}
+# INTEL: vfmaddcph xmm6 {k7}, xmm5, xmmword ptr [esp + 8*esi + 268435456]
+0x62,0xf6,0x56,0x0f,0x56,0xb4,0xf4,0x00,0x00,0x00,0x10
+
+# ATT:   vfmaddcph  (%ecx){1to4}, %xmm5, %xmm6
+# INTEL: vfmaddcph xmm6, xmm5, dword ptr [ecx]{1to4}
+0x62,0xf6,0x56,0x18,0x56,0x31
+
+# ATT:   vfmaddcph  2032(%ecx), %xmm5, %xmm6
+# INTEL: vfmaddcph xmm6, xmm5, xmmword ptr [ecx + 2032]
+0x62,0xf6,0x56,0x08,0x56,0x71,0x7f
+
+# ATT:   vfmaddcph  -512(%edx){1to4}, %xmm5, %xmm6 {%k7} {z}
+# INTEL: vfmaddcph xmm6 {k7} {z}, xmm5, dword ptr [edx - 512]{1to4}
+0x62,0xf6,0x56,0x9f,0x56,0x72,0x80
+
+# ATT:   vfmulcph %ymm4, %ymm5, %ymm6
+# INTEL: vfmulcph ymm6, ymm5, ymm4
+0x62,0xf6,0x56,0x28,0xd6,0xf4
+
+# ATT:   vfmulcph %xmm4, %xmm5, %xmm6
+# INTEL: vfmulcph xmm6, xmm5, xmm4
+0x62,0xf6,0x56,0x08,0xd6,0xf4
+
+# ATT:   vfmulcph  268435456(%esp,%esi,8), %ymm5, %ymm6 {%k7}
+# INTEL: vfmulcph ymm6 {k7}, ymm5, ymmword ptr [esp + 8*esi + 268435456]
+0x62,0xf6,0x56,0x2f,0xd6,0xb4,0xf4,0x00,0x00,0x00,0x10
+
+# ATT:   vfmulcph  (%ecx){1to8}, %ymm5, %ymm6
+# INTEL: vfmulcph ymm6, ymm5, dword ptr [ecx]{1to8}
+0x62,0xf6,0x56,0x38,0xd6,0x31
+
+# ATT:   vfmulcph  4064(%ecx), %ymm5, %ymm6
+# INTEL: vfmulcph ymm6, ymm5, ymmword ptr [ecx + 4064]
+0x62,0xf6,0x56,0x28,0xd6,0x71,0x7f
+
+# ATT:   vfmulcph  -512(%edx){1to8}, %ymm5, %ymm6 {%k7} {z}
+# INTEL: vfmulcph ymm6 {k7} {z}, ymm5, dword ptr [edx - 512]{1to8}
+0x62,0xf6,0x56,0xbf,0xd6,0x72,0x80
+
+# ATT:   vfmulcph  268435456(%esp,%esi,8), %xmm5, %xmm6 {%k7}
+# INTEL: vfmulcph xmm6 {k7}, xmm5, xmmword ptr [esp + 8*esi + 268435456]
+0x62,0xf6,0x56,0x0f,0xd6,0xb4,0xf4,0x00,0x00,0x00,0x10
+
+# ATT:   vfmulcph  (%ecx){1to4}, %xmm5, %xmm6
+# INTEL: vfmulcph xmm6, xmm5, dword ptr [ecx]{1to4}
+0x62,0xf6,0x56,0x18,0xd6,0x31
+
+# ATT:   vfmulcph  2032(%ecx), %xmm5, %xmm6
+# INTEL: vfmulcph xmm6, xmm5, xmmword ptr [ecx + 2032]
+0x62,0xf6,0x56,0x08,0xd6,0x71,0x7f
+
+# ATT:   vfmulcph  -512(%edx){1to4}, %xmm5, %xmm6 {%k7} {z}
+# INTEL: vfmulcph xmm6 {k7} {z}, xmm5, dword ptr [edx - 512]{1to4}
+0x62,0xf6,0x56,0x9f,0xd6,0x72,0x80

llvm/test/MC/X86/avx512fp16-complex-fma.s

+// RUN: llvm-mc -triple x86_64-unknown-unknown %s > %t 2> %t.err
+// RUN: FileCheck < %t %s
+// RUN: FileCheck --check-prefix=CHECK-STDERR < %t.err %s
+
+// CHECK: vfcmaddcph %zmm24, %zmm23, %zmm24
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph %zmm24, %zmm23, %zmm24
+
+// CHECK: vfcmaddcph {rn-sae}, %zmm24, %zmm23, %zmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph {rn-sae}, %zmm24, %zmm23, %zmm23
+
+// CHECK: vfcmaddcph %zmm24, %zmm23, %zmm24 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph %zmm24, %zmm23, %zmm24 {%k7}
+
+// CHECK: vfcmaddcph {rz-sae}, %zmm24, %zmm23, %zmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph {rz-sae}, %zmm24, %zmm23, %zmm23 {%k7} {z}
+
+// CHECK: vfcmaddcph  268435456(%rbp,%r14,8), %zmm23, %zmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph  268435456(%rbp,%r14,8), %zmm23, %zmm23
+
+// CHECK: vfcmaddcph  291(%r8,%rax,4), %zmm23, %zmm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph  291(%r8,%rax,4), %zmm23, %zmm23 {%k7}
+
+// CHECK: vfcmaddcph  (%rip){1to16}, %zmm23, %zmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph  (%rip){1to16}, %zmm23, %zmm23
+
+// CHECK: vfcmaddcph  -2048(,%rbp,2), %zmm23, %zmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph  -2048(,%rbp,2), %zmm23, %zmm23
+
+// CHECK: vfcmaddcph  8128(%rcx), %zmm23, %zmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph  8128(%rcx), %zmm23, %zmm23 {%k7} {z}
+
+// CHECK: vfcmaddcph  -512(%rdx){1to16}, %zmm23, %zmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph  -512(%rdx){1to16}, %zmm23, %zmm23 {%k7} {z}
+
+// CHECK: vfcmaddcsh %xmm24, %xmm23, %xmm24
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcsh %xmm24, %xmm23, %xmm24
+
+// CHECK: vfcmaddcsh {rn-sae}, %xmm24, %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcsh {rn-sae}, %xmm24, %xmm23, %xmm23
+
+// CHECK: vfcmaddcsh %xmm24, %xmm23, %xmm24 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcsh %xmm24, %xmm23, %xmm24 {%k7}
+
+// CHECK: vfcmaddcsh {rz-sae}, %xmm24, %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcsh {rz-sae}, %xmm24, %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfcmaddcsh  268435456(%rbp,%r14,8), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcsh  268435456(%rbp,%r14,8), %xmm23, %xmm23
+
+// CHECK: vfcmaddcsh  291(%r8,%rax,4), %xmm23, %xmm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcsh  291(%r8,%rax,4), %xmm23, %xmm23 {%k7}
+
+// CHECK: vfcmaddcsh  (%rip), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcsh  (%rip), %xmm23, %xmm23
+
+// CHECK: vfcmaddcsh  -128(,%rbp,2), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcsh  -128(,%rbp,2), %xmm23, %xmm23
+
+// CHECK: vfcmaddcsh  508(%rcx), %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcsh  508(%rcx), %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfcmaddcsh  -512(%rdx), %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcsh  -512(%rdx), %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfcmulcph %zmm24, %zmm23, %zmm24
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph %zmm24, %zmm23, %zmm24
+
+// CHECK: vfcmulcph {rn-sae}, %zmm24, %zmm23, %zmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph {rn-sae}, %zmm24, %zmm23, %zmm23
+
+// CHECK: vfcmulcph %zmm24, %zmm23, %zmm24 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph %zmm24, %zmm23, %zmm24 {%k7}
+
+// CHECK: vfcmulcph {rz-sae}, %zmm24, %zmm23, %zmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph {rz-sae}, %zmm24, %zmm23, %zmm23 {%k7} {z}
+
+// CHECK: vfcmulcph  268435456(%rbp,%r14,8), %zmm23, %zmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph  268435456(%rbp,%r14,8), %zmm23, %zmm23
+
+// CHECK: vfcmulcph  291(%r8,%rax,4), %zmm23, %zmm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph  291(%r8,%rax,4), %zmm23, %zmm23 {%k7}
+
+// CHECK: vfcmulcph  (%rip){1to16}, %zmm23, %zmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph  (%rip){1to16}, %zmm23, %zmm23
+
+// CHECK: vfcmulcph  -2048(,%rbp,2), %zmm23, %zmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph  -2048(,%rbp,2), %zmm23, %zmm23
+
+// CHECK: vfcmulcph  8128(%rcx), %zmm23, %zmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph  8128(%rcx), %zmm23, %zmm23 {%k7} {z}
+
+// CHECK: vfcmulcph  -512(%rdx){1to16}, %zmm23, %zmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph  -512(%rdx){1to16}, %zmm23, %zmm23 {%k7} {z}
+
+// CHECK: vfcmulcsh %xmm24, %xmm23, %xmm24
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcsh %xmm24, %xmm23, %xmm24
+
+// CHECK: vfcmulcsh {rn-sae}, %xmm24, %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcsh {rn-sae}, %xmm24, %xmm23, %xmm23
+
+// CHECK: vfcmulcsh %xmm24, %xmm23, %xmm24 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcsh %xmm24, %xmm23, %xmm24 {%k7}
+
+// CHECK: vfcmulcsh {rz-sae}, %xmm24, %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcsh {rz-sae}, %xmm24, %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfcmulcsh  268435456(%rbp,%r14,8), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcsh  268435456(%rbp,%r14,8), %xmm23, %xmm23
+
+// CHECK: vfcmulcsh  291(%r8,%rax,4), %xmm23, %xmm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcsh  291(%r8,%rax,4), %xmm23, %xmm23 {%k7}
+
+// CHECK: vfcmulcsh  (%rip), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcsh  (%rip), %xmm23, %xmm23
+
+// CHECK: vfcmulcsh  -128(,%rbp,2), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcsh  -128(,%rbp,2), %xmm23, %xmm23
+
+// CHECK: vfcmulcsh  508(%rcx), %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcsh  508(%rcx), %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfcmulcsh  -512(%rdx), %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcsh  -512(%rdx), %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfmaddcph %zmm24, %zmm23, %zmm24
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph %zmm24, %zmm23, %zmm24
+
+// CHECK: vfmaddcph {rn-sae}, %zmm24, %zmm23, %zmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph {rn-sae}, %zmm24, %zmm23, %zmm23
+
+// CHECK: vfmaddcph %zmm24, %zmm23, %zmm24 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph %zmm24, %zmm23, %zmm24 {%k7}
+
+// CHECK: vfmaddcph {rz-sae}, %zmm24, %zmm23, %zmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph {rz-sae}, %zmm24, %zmm23, %zmm23 {%k7} {z}
+
+// CHECK: vfmaddcph  268435456(%rbp,%r14,8), %zmm23, %zmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph  268435456(%rbp,%r14,8), %zmm23, %zmm23
+
+// CHECK: vfmaddcph  291(%r8,%rax,4), %zmm23, %zmm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph  291(%r8,%rax,4), %zmm23, %zmm23 {%k7}
+
+// CHECK: vfmaddcph  (%rip){1to16}, %zmm23, %zmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph  (%rip){1to16}, %zmm23, %zmm23
+
+// CHECK: vfmaddcph  -2048(,%rbp,2), %zmm23, %zmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph  -2048(,%rbp,2), %zmm23, %zmm23
+
+// CHECK: vfmaddcph  8128(%rcx), %zmm23, %zmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph  8128(%rcx), %zmm23, %zmm23 {%k7} {z}
+
+// CHECK: vfmaddcph  -512(%rdx){1to16}, %zmm23, %zmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph  -512(%rdx){1to16}, %zmm23, %zmm23 {%k7} {z}
+
+// CHECK: vfmaddcsh %xmm24, %xmm23, %xmm24
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcsh %xmm24, %xmm23, %xmm24
+
+// CHECK: vfmaddcsh {rn-sae}, %xmm24, %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcsh {rn-sae}, %xmm24, %xmm23, %xmm23
+
+// CHECK: vfmaddcsh %xmm24, %xmm23, %xmm24 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcsh %xmm24, %xmm23, %xmm24 {%k7}
+
+// CHECK: vfmaddcsh {rz-sae}, %xmm24, %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcsh {rz-sae}, %xmm24, %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfmaddcsh  268435456(%rbp,%r14,8), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcsh  268435456(%rbp,%r14,8), %xmm23, %xmm23
+
+// CHECK: vfmaddcsh  291(%r8,%rax,4), %xmm23, %xmm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcsh  291(%r8,%rax,4), %xmm23, %xmm23 {%k7}
+
+// CHECK: vfmaddcsh  (%rip), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcsh  (%rip), %xmm23, %xmm23
+
+// CHECK: vfmaddcsh  -128(,%rbp,2), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcsh  -128(,%rbp,2), %xmm23, %xmm23
+
+// CHECK: vfmaddcsh  508(%rcx), %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcsh  508(%rcx), %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfmaddcsh  -512(%rdx), %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcsh  -512(%rdx), %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfmulcph %zmm24, %zmm23, %zmm24
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph %zmm24, %zmm23, %zmm24
+
+// CHECK: vfmulcph {rn-sae}, %zmm24, %zmm23, %zmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph {rn-sae}, %zmm24, %zmm23, %zmm23
+
+// CHECK: vfmulcph %zmm24, %zmm23, %zmm24 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph %zmm24, %zmm23, %zmm24 {%k7}
+
+// CHECK: vfmulcph {rz-sae}, %zmm24, %zmm23, %zmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph {rz-sae}, %zmm24, %zmm23, %zmm23 {%k7} {z}
+
+// CHECK: vfmulcph  268435456(%rbp,%r14,8), %zmm23, %zmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph  268435456(%rbp,%r14,8), %zmm23, %zmm23
+
+// CHECK: vfmulcph  291(%r8,%rax,4), %zmm23, %zmm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph  291(%r8,%rax,4), %zmm23, %zmm23 {%k7}
+
+// CHECK: vfmulcph  (%rip){1to16}, %zmm23, %zmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph  (%rip){1to16}, %zmm23, %zmm23
+
+// CHECK: vfmulcph  -2048(,%rbp,2), %zmm23, %zmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph  -2048(,%rbp,2), %zmm23, %zmm23
+
+// CHECK: vfmulcph  8128(%rcx), %zmm23, %zmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph  8128(%rcx), %zmm23, %zmm23 {%k7} {z}
+
+// CHECK: vfmulcph  -512(%rdx){1to16}, %zmm23, %zmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph  -512(%rdx){1to16}, %zmm23, %zmm23 {%k7} {z}
+
+// CHECK: vfmulcsh %xmm24, %xmm23, %xmm24
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcsh %xmm24, %xmm23, %xmm24
+
+// CHECK: vfmulcsh {rn-sae}, %xmm24, %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcsh {rn-sae}, %xmm24, %xmm23, %xmm23
+
+// CHECK: vfmulcsh %xmm24, %xmm23, %xmm24 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcsh %xmm24, %xmm23, %xmm24 {%k7}
+
+// CHECK: vfmulcsh {rz-sae}, %xmm24, %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcsh {rz-sae}, %xmm24, %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfmulcsh  268435456(%rbp,%r14,8), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcsh  268435456(%rbp,%r14,8), %xmm23, %xmm23
+
+// CHECK: vfmulcsh  291(%r8,%rax,4), %xmm23, %xmm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcsh  291(%r8,%rax,4), %xmm23, %xmm23 {%k7}
+
+// CHECK: vfmulcsh  (%rip), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcsh  (%rip), %xmm23, %xmm23
+
+// CHECK: vfmulcsh  -128(,%rbp,2), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcsh  -128(,%rbp,2), %xmm23, %xmm23
+
+// CHECK: vfmulcsh  508(%rcx), %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcsh  508(%rcx), %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfmulcsh  -512(%rdx), %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcsh  -512(%rdx), %xmm23, %xmm23 {%k7} {z}
+

llvm/test/MC/X86/avx512fp16-complex-fma_vl.s

+// RUN: llvm-mc -triple x86_64-unknown-unknown %s > %t 2> %t.err
+// RUN: FileCheck < %t %s
+// RUN: FileCheck --check-prefix=CHECK-STDERR < %t.err %s
+
+// CHECK: vfcmaddcph %ymm24, %ymm23, %ymm24
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph %ymm24, %ymm23, %ymm24
+
+// CHECK: vfcmaddcph %ymm24, %ymm23, %ymm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph %ymm24, %ymm23, %ymm23 {%k7}
+
+// CHECK: vfcmaddcph %ymm24, %ymm23, %ymm24 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph %ymm24, %ymm23, %ymm24 {%k7} {z}
+
+// CHECK: vfcmaddcph %xmm24, %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph %xmm24, %xmm23, %xmm23
+
+// CHECK: vfcmaddcph %xmm24, %xmm23, %xmm24 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph %xmm24, %xmm23, %xmm24 {%k7}
+
+// CHECK: vfcmaddcph %xmm24, %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph %xmm24, %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfcmaddcph  268435456(%rbp,%r14,8), %ymm23, %ymm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph  268435456(%rbp,%r14,8), %ymm23, %ymm23
+
+// CHECK: vfcmaddcph  291(%r8,%rax,4), %ymm23, %ymm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph  291(%r8,%rax,4), %ymm23, %ymm23 {%k7}
+
+// CHECK: vfcmaddcph  (%rip){1to8}, %ymm23, %ymm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph  (%rip){1to8}, %ymm23, %ymm23
+
+// CHECK: vfcmaddcph  -1024(,%rbp,2), %ymm23, %ymm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph  -1024(,%rbp,2), %ymm23, %ymm23
+
+// CHECK: vfcmaddcph  4064(%rcx), %ymm23, %ymm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph  4064(%rcx), %ymm23, %ymm23 {%k7} {z}
+
+// CHECK: vfcmaddcph  -512(%rdx){1to8}, %ymm23, %ymm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph  -512(%rdx){1to8}, %ymm23, %ymm23 {%k7} {z}
+
+// CHECK: vfcmaddcph  268435456(%rbp,%r14,8), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph  268435456(%rbp,%r14,8), %xmm23, %xmm23
+
+// CHECK: vfcmaddcph  291(%r8,%rax,4), %xmm23, %xmm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph  291(%r8,%rax,4), %xmm23, %xmm23 {%k7}
+
+// CHECK: vfcmaddcph  (%rip){1to4}, %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph  (%rip){1to4}, %xmm23, %xmm23
+
+// CHECK: vfcmaddcph  -512(,%rbp,2), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph  -512(,%rbp,2), %xmm23, %xmm23
+
+// CHECK: vfcmaddcph  2032(%rcx), %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph  2032(%rcx), %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfcmaddcph  -512(%rdx){1to4}, %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmaddcph  -512(%rdx){1to4}, %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfcmulcph %ymm24, %ymm23, %ymm24
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph %ymm24, %ymm23, %ymm24
+
+// CHECK: vfcmulcph %ymm24, %ymm23, %ymm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph %ymm24, %ymm23, %ymm23 {%k7}
+
+// CHECK: vfcmulcph %ymm24, %ymm23, %ymm24 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph %ymm24, %ymm23, %ymm24 {%k7} {z}
+
+// CHECK: vfcmulcph %xmm24, %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph %xmm24, %xmm23, %xmm23
+
+// CHECK: vfcmulcph %xmm24, %xmm23, %xmm24 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph %xmm24, %xmm23, %xmm24 {%k7}
+
+// CHECK: vfcmulcph %xmm24, %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph %xmm24, %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfcmulcph  268435456(%rbp,%r14,8), %ymm23, %ymm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph  268435456(%rbp,%r14,8), %ymm23, %ymm23
+
+// CHECK: vfcmulcph  291(%r8,%rax,4), %ymm23, %ymm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph  291(%r8,%rax,4), %ymm23, %ymm23 {%k7}
+
+// CHECK: vfcmulcph  (%rip){1to8}, %ymm23, %ymm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph  (%rip){1to8}, %ymm23, %ymm23
+
+// CHECK: vfcmulcph  -1024(,%rbp,2), %ymm23, %ymm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph  -1024(,%rbp,2), %ymm23, %ymm23
+
+// CHECK: vfcmulcph  4064(%rcx), %ymm23, %ymm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph  4064(%rcx), %ymm23, %ymm23 {%k7} {z}
+
+// CHECK: vfcmulcph  -512(%rdx){1to8}, %ymm23, %ymm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph  -512(%rdx){1to8}, %ymm23, %ymm23 {%k7} {z}
+
+// CHECK: vfcmulcph  268435456(%rbp,%r14,8), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph  268435456(%rbp,%r14,8), %xmm23, %xmm23
+
+// CHECK: vfcmulcph  291(%r8,%rax,4), %xmm23, %xmm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph  291(%r8,%rax,4), %xmm23, %xmm23 {%k7}
+
+// CHECK: vfcmulcph  (%rip){1to4}, %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph  (%rip){1to4}, %xmm23, %xmm23
+
+// CHECK: vfcmulcph  -512(,%rbp,2), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph  -512(,%rbp,2), %xmm23, %xmm23
+
+// CHECK: vfcmulcph  2032(%rcx), %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph  2032(%rcx), %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfcmulcph  -512(%rdx){1to4}, %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfcmulcph  -512(%rdx){1to4}, %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfmaddcph %ymm24, %ymm23, %ymm24
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph %ymm24, %ymm23, %ymm24
+
+// CHECK: vfmaddcph %ymm24, %ymm23, %ymm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph %ymm24, %ymm23, %ymm23 {%k7}
+
+// CHECK: vfmaddcph %ymm24, %ymm23, %ymm24 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph %ymm24, %ymm23, %ymm24 {%k7} {z}
+
+// CHECK: vfmaddcph %xmm24, %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph %xmm24, %xmm23, %xmm23
+
+// CHECK: vfmaddcph %xmm24, %xmm23, %xmm24 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph %xmm24, %xmm23, %xmm24 {%k7}
+
+// CHECK: vfmaddcph %xmm24, %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph %xmm24, %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfmaddcph  268435456(%rbp,%r14,8), %ymm23, %ymm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph  268435456(%rbp,%r14,8), %ymm23, %ymm23
+
+// CHECK: vfmaddcph  291(%r8,%rax,4), %ymm23, %ymm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph  291(%r8,%rax,4), %ymm23, %ymm23 {%k7}
+
+// CHECK: vfmaddcph  (%rip){1to8}, %ymm23, %ymm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph  (%rip){1to8}, %ymm23, %ymm23
+
+// CHECK: vfmaddcph  -1024(,%rbp,2), %ymm23, %ymm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph  -1024(,%rbp,2), %ymm23, %ymm23
+
+// CHECK: vfmaddcph  4064(%rcx), %ymm23, %ymm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph  4064(%rcx), %ymm23, %ymm23 {%k7} {z}
+
+// CHECK: vfmaddcph  -512(%rdx){1to8}, %ymm23, %ymm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph  -512(%rdx){1to8}, %ymm23, %ymm23 {%k7} {z}
+
+// CHECK: vfmaddcph  268435456(%rbp,%r14,8), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph  268435456(%rbp,%r14,8), %xmm23, %xmm23
+
+// CHECK: vfmaddcph  291(%r8,%rax,4), %xmm23, %xmm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph  291(%r8,%rax,4), %xmm23, %xmm23 {%k7}
+
+// CHECK: vfmaddcph  (%rip){1to4}, %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph  (%rip){1to4}, %xmm23, %xmm23
+
+// CHECK: vfmaddcph  -512(,%rbp,2), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph  -512(,%rbp,2), %xmm23, %xmm23
+
+// CHECK: vfmaddcph  2032(%rcx), %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph  2032(%rcx), %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfmaddcph  -512(%rdx){1to4}, %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmaddcph  -512(%rdx){1to4}, %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfmulcph %ymm24, %ymm23, %ymm24
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph %ymm24, %ymm23, %ymm24
+
+// CHECK: vfmulcph %ymm24, %ymm23, %ymm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph %ymm24, %ymm23, %ymm23 {%k7}
+
+// CHECK: vfmulcph %ymm24, %ymm23, %ymm24 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph %ymm24, %ymm23, %ymm24 {%k7} {z}
+
+// CHECK: vfmulcph %xmm24, %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph %xmm24, %xmm23, %xmm23
+
+// CHECK: vfmulcph %xmm24, %xmm23, %xmm24 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph %xmm24, %xmm23, %xmm24 {%k7}
+
+// CHECK: vfmulcph %xmm24, %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph %xmm24, %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfmulcph  268435456(%rbp,%r14,8), %ymm23, %ymm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph  268435456(%rbp,%r14,8), %ymm23, %ymm23
+
+// CHECK: vfmulcph  291(%r8,%rax,4), %ymm23, %ymm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph  291(%r8,%rax,4), %ymm23, %ymm23 {%k7}
+
+// CHECK: vfmulcph  (%rip){1to8}, %ymm23, %ymm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph  (%rip){1to8}, %ymm23, %ymm23
+
+// CHECK: vfmulcph  -1024(,%rbp,2), %ymm23, %ymm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph  -1024(,%rbp,2), %ymm23, %ymm23
+
+// CHECK: vfmulcph  4064(%rcx), %ymm23, %ymm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph  4064(%rcx), %ymm23, %ymm23 {%k7} {z}
+
+// CHECK: vfmulcph  -512(%rdx){1to8}, %ymm23, %ymm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph  -512(%rdx){1to8}, %ymm23, %ymm23 {%k7} {z}
+
+// CHECK: vfmulcph  268435456(%rbp,%r14,8), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph  268435456(%rbp,%r14,8), %xmm23, %xmm23
+
+// CHECK: vfmulcph  291(%r8,%rax,4), %xmm23, %xmm23 {%k7}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph  291(%r8,%rax,4), %xmm23, %xmm23 {%k7}
+
+// CHECK: vfmulcph  (%rip){1to4}, %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph  (%rip){1to4}, %xmm23, %xmm23
+
+// CHECK: vfmulcph  -512(,%rbp,2), %xmm23, %xmm23
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph  -512(,%rbp,2), %xmm23, %xmm23
+
+// CHECK: vfmulcph  2032(%rcx), %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph  2032(%rcx), %xmm23, %xmm23 {%k7} {z}
+
+// CHECK: vfmulcph  -512(%rdx){1to4}, %xmm23, %xmm23 {%k7} {z}
+// CHECK-STDERR: warning: Destination register should be distinct from source registers
+          vfmulcph  -512(%rdx){1to4}, %xmm23, %xmm23 {%k7} {z}
+

llvm/test/MC/X86/avx512fp16.s

 
 // CHECK: vfnmsub231sh  254(%rcx), %xmm29, %xmm30
 // CHECK: encoding: [0x62,0x66,0x15,0x00,0xbf,0x71,0x7f]
           vfnmsub231sh  254(%rcx), %xmm29, %xmm30
 
 // CHECK: vfnmsub231sh  -256(%rdx), %xmm29, %xmm30 {%k7} {z}
 // CHECK: encoding: [0x62,0x66,0x15,0x87,0xbf,0x72,0x80]
           vfnmsub231sh  -256(%rdx), %xmm29, %xmm30 {%k7} {z}
+
+// CHECK: vfcmaddcph %zmm28, %zmm29, %zmm30
+// CHECK: encoding: [0x62,0x06,0x17,0x40,0x56,0xf4]
+          vfcmaddcph %zmm28, %zmm29, %zmm30
+
+// CHECK: vfcmaddcph {rn-sae}, %zmm28, %zmm29, %zmm30
+// CHECK: encoding: [0x62,0x06,0x17,0x10,0x56,0xf4]
+          vfcmaddcph {rn-sae}, %zmm28, %zmm29, %zmm30
+
+// CHECK: vfcmaddcph  268435456(%rbp,%r14,8), %zmm29, %zmm30 {%k7}
+// CHECK: encoding: [0x62,0x26,0x17,0x47,0x56,0xb4,0xf5,0x00,0x00,0x00,0x10]
+          vfcmaddcph  268435456(%rbp,%r14,8), %zmm29, %zmm30 {%k7}
+
+// CHECK: vfcmaddcph  (%r9){1to16}, %zmm29, %zmm30
+// CHECK: encoding: [0x62,0x46,0x17,0x50,0x56,0x31]
+          vfcmaddcph  (%r9){1to16}, %zmm29, %zmm30
+
+// CHECK: vfcmaddcph  8128(%rcx), %zmm29, %zmm30
+// CHECK: encoding: [0x62,0x66,0x17,0x40,0x56,0x71,0x7f]
+          vfcmaddcph  8128(%rcx), %zmm29, %zmm30
+
+// CHECK: vfcmaddcph  -512(%rdx){1to16}, %zmm29, %zmm30 {%k7} {z}
+// CHECK: encoding: [0x62,0x66,0x17,0xd7,0x56,0x72,0x80]
+          vfcmaddcph  -512(%rdx){1to16}, %zmm29, %zmm30 {%k7} {z}
+
+// CHECK: vfcmaddcsh %xmm28, %xmm29, %xmm30
+// CHECK: encoding: [0x62,0x06,0x17,0x00,0x57,0xf4]
+          vfcmaddcsh %xmm28, %xmm29, %xmm30
+
+// CHECK: vfcmaddcsh {rn-sae}, %xmm28, %xmm29, %xmm30
+// CHECK: encoding: [0x62,0x06,0x17,0x10,0x57,0xf4]
+          vfcmaddcsh {rn-sae}, %xmm28, %xmm29, %xmm30
+
+// CHECK: vfcmaddcsh  268435456(%rbp,%r14,8), %xmm29, %xmm30 {%k7}
+// CHECK: encoding: [0x62,0x26,0x17,0x07,0x57,0xb4,0xf5,0x00,0x00,0x00,0x10]
+          vfcmaddcsh  268435456(%rbp,%r14,8), %xmm29, %xmm30 {%k7}
+
+// CHECK: vfcmaddcsh  (%r9), %xmm29, %xmm30
+// CHECK: encoding: [0x62,0x46,0x17,0x00,0x57,0x31]
+          vfcmaddcsh  (%r9), %xmm29, %xmm30
+
+// CHECK: vfcmaddcsh  508(%rcx), %xmm29, %xmm30
+// CHECK: encoding: [0x62,0x66,0x17,0x00,0x57,0x71,0x7f]
+          vfcmaddcsh  508(%rcx), %xmm29, %xmm30
+
+// CHECK: vfcmaddcsh  -512(%rdx), %xmm29, %xmm30 {%k7} {z}
+// CHECK: encoding: [0x62,0x66,0x17,0x87,0x57,0x72,0x80]
+          vfcmaddcsh  -512(%rdx), %xmm29, %xmm30 {%k7} {z}
+
+// CHECK: vfcmulcph %zmm28, %zmm29, %zmm30
+// CHECK: encoding: [0x62,0x06,0x17,0x40,0xd6,0xf4]
+          vfcmulcph %zmm28, %zmm29, %zmm30
+
+// CHECK: vfcmulcph {rn-sae}, %zmm28, %zmm29, %zmm30
+// CHECK: encoding: [0x62,0x06,0x17,0x10,0xd6,0xf4]
+          vfcmulcph {rn-sae}, %zmm28, %zmm29, %zmm30
+
+// CHECK: vfcmulcph  268435456(%rbp,%r14,8), %zmm29, %zmm30 {%k7}
+// CHECK: encoding: [0x62,0x26,0x17,0x47,0xd6,0xb4,0xf5,0x00,0x00,0x00,0x10]
+          vfcmulcph  268435456(%rbp,%r14,8), %zmm29, %zmm30 {%k7}
+
+// CHECK: vfcmulcph  (%r9){1to16}, %zmm29, %zmm30
+// CHECK: encoding: [0x62,0x46,0x17,0x50,0xd6,0x31]
+          vfcmulcph  (%r9){1to16}, %zmm29, %zmm30
+
+// CHECK: vfcmulcph  8128(%rcx), %zmm29, %zmm30
+// CHECK: encoding: [0x62,0x66,0x17,0x40,0xd6,0x71,0x7f]
+          vfcmulcph  8128(%rcx), %zmm29, %zmm30
+
+// CHECK: vfcmulcph  -512(%rdx){1to16}, %zmm29, %zmm30 {%k7} {z}
+// CHECK: encoding: [0x62,0x66,0x17,0xd7,0xd6,0x72,0x80]
+          vfcmulcph  -512(%rdx){1to16}, %zmm29, %zmm30 {%k7} {z}
+
+// CHECK: vfcmulcsh %xmm28, %xmm29, %xmm30
+// CHECK: encoding: [0x62,0x06,0x17,0x00,0xd7,0xf4]
+          vfcmulcsh %xmm28, %xmm29, %xmm30
+
+// CHECK: vfcmulcsh {rn-sae}, %xmm28, %xmm29, %xmm30
+// CHECK: encoding: [0x62,0x06,0x17,0x10,0xd7,0xf4]
+          vfcmulcsh {rn-sae}, %xmm28, %xmm29, %xmm30
+
+// CHECK: vfcmulcsh  268435456(%rbp,%r14,8), %xmm29, %xmm30 {%k7}
+// CHECK: encoding: [0x62,0x26,0x17,0x07,0xd7,0xb4,0xf5,0x00,0x00,0x00,0x10]
+          vfcmulcsh  268435456(%rbp,%r14,8), %xmm29, %xmm30 {%k7}
+
+// CHECK: vfcmulcsh  (%r9), %xmm29, %xmm30
+// CHECK: encoding: [0x62,0x46,0x17,0x00,0xd7,0x31]
+          vfcmulcsh  (%r9), %xmm29, %xmm30
+
+// CHECK: vfcmulcsh  508(%rcx), %xmm29, %xmm30
+// CHECK: encoding: [0x62,0x66,0x17,0x00,0xd7,0x71,0x7f]
+          vfcmulcsh  508(%rcx), %xmm29, %xmm30
+
+// CHECK: vfcmulcsh  -512(%rdx), %xmm29, %xmm30 {%k7} {z}
+// CHECK: encoding: [0x62,0x66,0x17,0x87,0xd7,0x72,0x80]
+          vfcmulcsh  -512(%rdx), %xmm29, %xmm30 {%k7} {z}
+
+// CHECK: vfmaddcph %zmm28, %zmm29, %zmm30
+// CHECK: encoding: [0x62,0x06,0x16,0x40,0x56,0xf4]
+          vfmaddcph %zmm28, %zmm29, %zmm30
+
+// CHECK: vfmaddcph {rn-sae}, %zmm28, %zmm29, %zmm30
+// CHECK: encoding: [0x62,0x06,0x16,0x10,0x56,0xf4]
+          vfmaddcph {rn-sae}, %zmm28, %zmm29, %zmm30
+
+// CHECK: vfmaddcph  268435456(%rbp,%r14,8), %zmm29, %zmm30 {%k7}
+// CHECK: encoding: [0x62,0x26,0x16,0x47,0x56,0xb4,0xf5,0x00,0x00,0x00,0x10]
+          vfmaddcph  268435456(%rbp,%r14,8), %zmm29, %zmm30 {%k7}
+
+// CHECK: vfmaddcph  (%r9){1to16}, %zmm29, %zmm30
+// CHECK: encoding: [0x62,0x46,0x16,0x50,0x56,0x31]
+          vfmaddcph  (%r9){1to16}, %zmm29, %zmm30
+
+// CHECK: vfmaddcph  8128(%rcx), %zmm29, %zmm30
+// CHECK: encoding: [0x62,0x66,0x16,0x40,0x56,0x71,0x7f]
+          vfmaddcph  8128(%rcx), %zmm29, %zmm30
+
+// CHECK: vfmaddcph  -512(%rdx){1to16}, %zmm29, %zmm30 {%k7} {z}
+// CHECK: encoding: [0x62,0x66,0x16,0xd7,0x56,0x72,0x80]
+          vfmaddcph  -512(%rdx){1to16}, %zmm29, %zmm30 {%k7} {z}
+
+// CHECK: vfmaddcsh %xmm28, %xmm29, %xmm30
+// CHECK: encoding: [0x62,0x06,0x16,0x00,0x57,0xf4]
+          vfmaddcsh %xmm28, %xmm29, %xmm30
+
+// CHECK: vfmaddcsh {rn-sae}, %xmm28, %xmm29, %xmm30
+// CHECK: encoding: [0x62,0x06,0x16,0x10,0x57,0xf4]
+          vfmaddcsh {rn-sae}, %xmm28, %xmm29, %xmm30
+
+// CHECK: vfmaddcsh  268435456(%rbp,%r14,8), %xmm29, %xmm30 {%k7}
+// CHECK: encoding: [0x62,0x26,0x16,0x07,0x57,0xb4,0xf5,0x00,0x00,0x00,0x10]
+          vfmaddcsh  268435456(%rbp,%r14,8), %xmm29, %xmm30 {%k7}
+
+// CHECK: vfmaddcsh  (%r9), %xmm29, %xmm30
+// CHECK: encoding: [0x62,0x46,0x16,0x00,0x57,0x31]
+          vfmaddcsh  (%r9), %xmm29, %xmm30
+
+// CHECK: vfmaddcsh  508(%rcx), %xmm29, %xmm30
+// CHECK: encoding: [0x62,0x66,0x16,0x00,0x57,0x71,0x7f]
+          vfmaddcsh  508(%rcx), %xmm29, %xmm30
+
+// CHECK: vfmaddcsh  -512(%rdx), %xmm29, %xmm30 {%k7} {z}
+// CHECK: encoding: [0x62,0x66,0x16,0x87,0x57,0x72,0x80]
+          vfmaddcsh  -512(%rdx), %xmm29, %xmm30 {%k7} {z}
+
+// CHECK: vfmulcph %zmm28, %zmm29, %zmm30
+// CHECK: encoding: [0x62,0x06,0x16,0x40,0xd6,0xf4]
+          vfmulcph %zmm28, %zmm29, %zmm30
+
+// CHECK: vfmulcph {rn-sae}, %zmm28, %zmm29, %zmm30
+// CHECK: encoding: [0x62,0x06,0x16,0x10,0xd6,0xf4]
+          vfmulcph {rn-sae}, %zmm28, %zmm29, %zmm30
+
+// CHECK: vfmulcph  268435456(%rbp,%r14,8), %zmm29, %zmm30 {%k7}
+// CHECK: encoding: [0x62,0x26,0x16,0x47,0xd6,0xb4,0xf5,0x00,0x00,0x00,0x10]
+          vfmulcph  268435456(%rbp,%r14,8), %zmm29, %zmm30 {%k7}
+
+// CHECK: vfmulcph  (%r9){1to16}, %zmm29, %zmm30
+// CHECK: encoding: [0x62,0x46,0x16,0x50,0xd6,0x31]
+          vfmulcph  (%r9){1to16}, %zmm29, %zmm30
+
+// CHECK: vfmulcph  8128(%rcx), %zmm29, %zmm30
+// CHECK: encoding: [0x62,0x66,0x16,0x40,0xd6,0x71,0x7f]
+          vfmulcph  8128(%rcx), %zmm29, %zmm30
+
+// CHECK: vfmulcph  -512(%rdx){1to16}, %zmm29, %zmm30 {%k7} {z}
+// CHECK: encoding: [0x62,0x66,0x16,0xd7,0xd6,0x72,0x80]
+          vfmulcph  -512(%rdx){1to16}, %zmm29, %zmm30 {%k7} {z}
+
+// CHECK: vfmulcsh %xmm28, %xmm29, %xmm30
+// CHECK: encoding: [0x62,0x06,0x16,0x00,0xd7,0xf4]
+          vfmulcsh %xmm28, %xmm29, %xmm30
+
+// CHECK: vfmulcsh {rn-sae}, %xmm28, %xmm29, %xmm30
+// CHECK: encoding: [0x62,0x06,0x16,0x10,0xd7,0xf4]
+          vfmulcsh {rn-sae}, %xmm28, %xmm29, %xmm30
+
+// CHECK: vfmulcsh  268435456(%rbp,%r14,8), %xmm29, %xmm30 {%k7}
+// CHECK: encoding: [0x62,0x26,0x16,0x07,0xd7,0xb4,0xf5,0x00,0x00,0x00,0x10]
+          vfmulcsh  268435456(%rbp,%r14,8), %xmm29, %xmm30 {%k7}
+
+// CHECK: vfmulcsh  (%r9), %xmm29, %xmm30
+// CHECK: encoding: [0x62,0x46,0x16,0x00,0xd7,0x31]
+          vfmulcsh  (%r9), %xmm29, %xmm30
+
+// CHECK: vfmulcsh  508(%rcx), %xmm29, %xmm30
+// CHECK: encoding: [0x62,0x66,0x16,0x00,0xd7,0x71,0x7f]
+          vfmulcsh  508(%rcx), %xmm29, %xmm30
+
+// CHECK: vfmulcsh  -512(%rdx), %xmm29, %xmm30 {%k7} {z}
+// CHECK: encoding: [0x62,0x66,0x16,0x87,0xd7,0x72,0x80]
+          vfmulcsh  -512(%rdx), %xmm29, %xmm30 {%k7} {z}

llvm/test/MC/X86/avx512fp16vl.s

 
 // CHECK: vfnmsub231ph  2032(%ecx), %xmm5, %xmm6
 // CHECK: encoding: [0x62,0xf6,0x55,0x08,0xbe,0x71,0x7f]
           vfnmsub231ph  2032(%ecx), %xmm5, %xmm6
 
 // CHECK: vfnmsub231ph  -256(%edx){1to8}, %xmm5, %xmm6 {%k7} {z}
 // CHECK: encoding: [0x62,0xf6,0x55,0x9f,0xbe,0x72,0x80]
           vfnmsub231ph  -256(%edx){1to8}, %xmm5, %xmm6 {%k7} {z}
+
+// CHECK: vfcmaddcph %ymm4, %ymm5, %ymm6
+// CHECK: encoding: [0x62,0xf6,0x57,0x28,0x56,0xf4]
+          vfcmaddcph %ymm4, %ymm5, %ymm6
+
+// CHECK: vfcmaddcph %xmm4, %xmm5, %xmm6
+// CHECK: encoding: [0x62,0xf6,0x57,0x08,0x56,0xf4]
+          vfcmaddcph %xmm4, %xmm5, %xmm6
+
+// CHECK: vfcmaddcph  268435456(%esp,%esi,8), %ymm5, %ymm6 {%k7}
+// CHECK: encoding: [0x62,0xf6,0x57,0x2f,0x56,0xb4,0xf4,0x00,0x00,0x00,0x10]
+          vfcmaddcph  268435456(%esp,%esi,8), %ymm5, %ymm6 {%k7}
+
+// CHECK: vfcmaddcph  (%ecx){1to8}, %ymm5, %ymm6
+// CHECK: encoding: [0x62,0xf6,0x57,0x38,0x56,0x31]
+          vfcmaddcph  (%ecx){1to8}, %ymm5, %ymm6
+
+// CHECK: vfcmaddcph  4064(%ecx), %ymm5, %ymm6
+// CHECK: encoding: [0x62,0xf6,0x57,0x28,0x56,0x71,0x7f]
+          vfcmaddcph  4064(%ecx), %ymm5, %ymm6
+
+// CHECK: vfcmaddcph  -512(%edx){1to8}, %ymm5, %ymm6 {%k7} {z}
+// CHECK: encoding: [0x62,0xf6,0x57,0xbf,0x56,0x72,0x80]
+          vfcmaddcph  -512(%edx){1to8}, %ymm5, %ymm6 {%k7} {z}
+
+// CHECK: vfcmaddcph  268435456(%esp,%esi,8), %xmm5, %xmm6 {%k7}
+// CHECK: encoding: [0x62,0xf6,0x57,0x0f,0x56,0xb4,0xf4,0x00,0x00,0x00,0x10]
+          vfcmaddcph  268435456(%esp,%esi,8), %xmm5, %xmm6 {%k7}
+
+// CHECK: vfcmaddcph  (%ecx){1to4}, %xmm5, %xmm6
+// CHECK: encoding: [0x62,0xf6,0x57,0x18,0x56,0x31]
+          vfcmaddcph  (%ecx){1to4}, %xmm5, %xmm6
+
+// CHECK: vfcmaddcph  2032(%ecx), %xmm5, %xmm6
+// CHECK: encoding: [0x62,0xf6,0x57,0x08,0x56,0x71,0x7f]
+          vfcmaddcph  2032(%ecx), %xmm5, %xmm6
+
+// CHECK: vfcmaddcph  -512(%edx){1to4}, %xmm5, %xmm6 {%k7} {z}
+// CHECK: encoding: [0x62,0xf6,0x57,0x9f,0x56,0x72,0x80]
+          vfcmaddcph  -512(%edx){1to4}, %xmm5, %xmm6 {%k7} {z}
+
+// CHECK: vfcmulcph %ymm4, %ymm5, %ymm6
+// CHECK: encoding: [0x62,0xf6,0x57,0x28,0xd6,0xf4]
+          vfcmulcph %ymm4, %ymm5, %ymm6
+
+// CHECK: vfcmulcph %xmm4, %xmm5, %xmm6
+// CHECK: encoding: [0x62,0xf6,0x57,0x08,0xd6,0xf4]
+          vfcmulcph %xmm4, %xmm5, %xmm6
+
+// CHECK: vfcmulcph  268435456(%esp,%esi,8), %ymm5, %ymm6 {%k7}
+// CHECK: encoding: [0x62,0xf6,0x57,0x2f,0xd6,0xb4,0xf4,0x00,0x00,0x00,0x10]
+          vfcmulcph  268435456(%esp,%esi,8), %ymm5, %ymm6 {%k7}
+
+// CHECK: vfcmulcph  (%ecx){1to8}, %ymm5, %ymm6
+// CHECK: encoding: [0x62,0xf6,0x57,0x38,0xd6,0x31]
+          vfcmulcph  (%ecx){1to8}, %ymm5, %ymm6
+
+// CHECK: vfcmulcph  4064(%ecx), %ymm5, %ymm6
+// CHECK: encoding: [0x62,0xf6,0x57,0x28,0xd6,0x71,0x7f]
+          vfcmulcph  4064(%ecx), %ymm5, %ymm6
+
+// CHECK: vfcmulcph  -512(%edx){1to8}, %ymm5, %ymm6 {%k7} {z}
+// CHECK: encoding: [0x62,0xf6,0x57,0xbf,0xd6,0x72,0x80]
+          vfcmulcph  -512(%edx){1to8}, %ymm5, %ymm6 {%k7} {z}
+
+// CHECK: vfcmulcph  268435456(%esp,%esi,8), %xmm5, %xmm6 {%k7}
+// CHECK: encoding: [0x62,0xf6,0x57,0x0f,0xd6,0xb4,0xf4,0x00,0x00,0x00,0x10]
+          vfcmulcph  268435456(%esp,%esi,8), %xmm5, %xmm6 {%k7}
+
+// CHECK: vfcmulcph  (%ecx){1to4}, %xmm5, %xmm6
+// CHECK: encoding: [0x62,0xf6,0x57,0x18,0xd6,0x31]
+          vfcmulcph  (%ecx){1to4}, %xmm5, %xmm6
+
+// CHECK: vfcmulcph  2032(%ecx), %xmm5, %xmm6
+// CHECK: encoding: [0x62,0xf6,0x57,0x08,0xd6,0x71,0x7f]
+          vfcmulcph  2032(%ecx), %xmm5, %xmm6
+
+// CHECK: vfcmulcph  -512(%edx){1to4}, %xmm5, %xmm6 {%k7} {z}
+// CHECK: encoding: [0x62,0xf6,0x57,0x9f,0xd6,0x72,0x80]
+          vfcmulcph  -512(%edx){1to4}, %xmm5, %xmm6 {%k7} {z}
+
+// CHECK: vfmaddcph %ymm4, %ymm5, %ymm6
+// CHECK: encoding: [0x62,0xf6,0x56,0x28,0x56,0xf4]
+          vfmaddcph %ymm4, %ymm5, %ymm6
+
+// CHECK: vfmaddcph %xmm4, %xmm5, %xmm6
+// CHECK: encoding: [0x62,0xf6,0x56,0x08,0x56,0xf4]
+          vfmaddcph %xmm4, %xmm5, %xmm6
+
+// CHECK: vfmaddcph  268435456(%esp,%esi,8), %ymm5, %ymm6 {%k7}
+// CHECK: encoding: [0x62,0xf6,0x56,0x2f,0x56,0xb4,0xf4,0x00,0x00,0x00,0x10]
+          vfmaddcph  268435456(%esp,%esi,8), %ymm5, %ymm6 {%k7}
+
+// CHECK: vfmaddcph  (%ecx){1to8}, %ymm5, %ymm6
+// CHECK: encoding: [0x62,0xf6,0x56,0x38,0x56,0x31]
+          vfmaddcph  (%ecx){1to8}, %ymm5, %ymm6
+
+// CHECK: vfmaddcph  4064(%ecx), %ymm5, %ymm6
+// CHECK: encoding: [0x62,0xf6,0x56,0x28,0x56,0x71,0x7f]
+          vfmaddcph  4064(%ecx), %ymm5, %ymm6
+
+// CHECK: vfmaddcph  -512(%edx){1to8}, %ymm5, %ymm6 {%k7} {z}
+// CHECK: encoding: [0x62,0xf6,0x56,0xbf,0x56,0x72,0x80]
+          vfmaddcph  -512(%edx){1to8}, %ymm5, %ymm6 {%k7} {z}
+
+// CHECK: vfmaddcph  268435456(%esp,%esi,8), %xmm5, %xmm6 {%k7}
+// CHECK: encoding: [0x62,0xf6,0x56,0x0f,0x56,0xb4,0xf4,0x00,0x00,0x00,0x10]
+          vfmaddcph  268435456(%esp,%esi,8), %xmm5, %xmm6 {%k7}
+
+// CHECK: vfmaddcph  (%ecx){1to4}, %xmm5, %xmm6
+// CHECK: encoding: [0x62,0xf6,0x56,0x18,0x56,0x31]
+          vfmaddcph  (%ecx){1to4}, %xmm5, %xmm6
+
+// CHECK: vfmaddcph  2032(%ecx), %xmm5, %xmm6
+// CHECK: encoding: [0x62,0xf6,0x56,0x08,0x56,0x71,0x7f]
+          vfmaddcph  2032(%ecx), %xmm5, %xmm6
+
+// CHECK: vfmaddcph  -512(%edx){1to4}, %xmm5, %xmm6 {%k7} {z}
+// CHECK: encoding: [0x62,0xf6,0x56,0x9f,0x56,0x72,0x80]
+          vfmaddcph  -512(%edx){1to4}, %xmm5, %xmm6 {%k7} {z}
+
+// CHECK: vfmulcph %ymm4, %ymm5, %ymm6
+// CHECK: encoding: [0x62,0xf6,0x56,0x28,0xd6,0xf4]
+          vfmulcph %ymm4, %ymm5, %ymm6
+
+// CHECK: vfmulcph %xmm4, %xmm5, %xmm6
+// CHECK: encoding: [0x62,0xf6,0x56,0x08,0xd6,0xf4]
+          vfmulcph %xmm4, %xmm5, %xmm6
+
+// CHECK: vfmulcph  268435456(%esp,%esi,8), %ymm5, %ymm6 {%k7}
+// CHECK: encoding: [0x62,0xf6,0x56,0x2f,0xd6,0xb4,0xf4,0x00,0x00,0x00,0x10]
+          vfmulcph  268435456(%esp,%esi,8), %ymm5, %ymm6 {%k7}
+
+// CHECK: vfmulcph  (%ecx){1to8}, %ymm5, %ymm6
+// CHECK: encoding: [0x62,0xf6,0x56,0x38,0xd6,0x31]
+          vfmulcph  (%ecx){1to8}, %ymm5, %ymm6
+
+// CHECK: vfmulcph  4064(%ecx), %ymm5, %ymm6
+// CHECK: encoding: [0x62,0xf6,0x56,0x28,0xd6,0x71,0x7f]
+          vfmulcph  4064(%ecx), %ymm5, %ymm6
+
+// CHECK: vfmulcph  -512(%edx){1to8}, %ymm5, %ymm6 {%k7} {z}
+// CHECK: encoding: [0x62,0xf6,0x56,0xbf,0xd6,0x72,0x80]
+          vfmulcph  -512(%edx){1to8}, %ymm5, %ymm6 {%k7} {z}
+
+// CHECK: vfmulcph  268435456(%esp,%esi,8), %xmm5, %xmm6 {%k7}
+// CHECK: encoding: [0x62,0xf6,0x56,0x0f,0xd6,0xb4,0xf4,0x00,0x00,0x00,0x10]
+          vfmulcph  268435456(%esp,%esi,8), %xmm5, %xmm6 {%k7}
+
+// CHECK: vfmulcph  (%ecx){1to4}, %xmm5, %xmm6
+// CHECK: encoding: [0x62,0xf6,0x56,0x18,0xd6,0x31]
+          vfmulcph  (%ecx){1to4}, %xmm5, %xmm6
+
+// CHECK: vfmulcph  2032(%ecx), %xmm5, %xmm6
+// CHECK: encoding: [0x62,0xf6,0x56,0x08,0xd6,0x71,0x7f]
+          vfmulcph  2032(%ecx), %xmm5, %xmm6
+
+// CHECK: vfmulcph  -512(%edx){1to4}, %xmm5, %xmm6 {%k7} {z}
+// CHECK: encoding: [0x62,0xf6,0x56,0x9f,0xd6,0x72,0x80]
+          vfmulcph  -512(%edx){1to4}, %xmm5, %xmm6 {%k7} {z}

llvm/test/MC/X86/intel-syntax-avx512fp16.s

 
 // CHECK: vfnmsub231sh xmm6, xmm5, word ptr [ecx + 254]
 // CHECK: encoding: [0x62,0xf6,0x55,0x08,0xbf,0x71,0x7f]
           vfnmsub231sh xmm6, xmm5, word ptr [ecx + 254]
 
 // CHECK: vfnmsub231sh xmm6 {k7} {z}, xmm5, word ptr [edx - 256]
 // CHECK: encoding: [0x62,0xf6,0x55,0x8f,0xbf,0x72,0x80]
           vfnmsub231sh xmm6 {k7} {z}, xmm5, word ptr [edx - 256]
+
+// CHECK: vfcmaddcph zmm6, zmm5, zmm4
+// CHECK: encoding: [0x62,0xf6,0x57,0x48,0x56,0xf4]
+          vfcmaddcph zmm6, zmm5, zmm4
+
+// CHECK: vfcmaddcph zmm6, zmm5, zmm4, {rn-sae}
+// CHECK: encoding: [0x62,0xf6,0x57,0x18,0x56,0xf4]
+          vfcmaddcph zmm6, zmm5, zmm4, {rn-sae}
+
+// CHECK: vfcmaddcph zmm6 {k7}, zmm5, zmmword ptr [esp + 8*esi + 268435456]
+// CHECK: encoding: [0x62,0xf6,0x57,0x4f,0x56,0xb4,0xf4,0x00,0x00,0x00,0x10]
+          vfcmaddcph zmm6 {k7}, zmm5, zmmword ptr [esp + 8*esi + 268435456]
+
+// CHECK: vfcmaddcph zmm6, zmm5, dword ptr [ecx]{1to16}
+// CHECK: encoding: [0x62,0xf6,0x57,0x58,0x56,0x31]
+          vfcmaddcph zmm6, zmm5, dword ptr [ecx]{1to16}
+
+// CHECK: vfcmaddcph zmm6, zmm5, zmmword ptr [ecx + 8128]
+// CHECK: encoding: [0x62,0xf6,0x57,0x48,0x56,0x71,0x7f]
+          vfcmaddcph zmm6, zmm5, zmmword ptr [ecx + 8128]
+
+// CHECK: vfcmaddcph zmm6 {k7} {z}, zmm5, dword ptr [edx - 512]{1to16}
+// CHECK: encoding: [0x62,0xf6,0x57,0xdf,0x56,0x72,0x80]
+          vfcmaddcph zmm6 {k7} {z}, zmm5, dword ptr [edx - 512]{1to16}
+
+// CHECK: vfcmaddcsh xmm6, xmm5, xmm4
+// CHECK: encoding: [0x62,0xf6,0x57,0x08,0x57,0xf4]
+          vfcmaddcsh xmm6, xmm5, xmm4
+
+// CHECK: vfcmaddcsh xmm6, xmm5, xmm4, {rn-sae}
+// CHECK: encoding: [0x62,0xf6,0x57,0x18,0x57,0xf4]
+          vfcmaddcsh xmm6, xmm5, xmm4, {rn-sae}
+
+// CHECK: vfcmaddcsh xmm6 {k7}, xmm5, dword ptr [esp + 8*esi + 268435456]
+// CHECK: encoding: [0x62,0xf6,0x57,0x0f,0x57,0xb4,0xf4,0x00,0x00,0x00,0x10]
+          vfcmaddcsh xmm6 {k7}, xmm5, dword ptr [esp + 8*esi + 268435456]
+
+// CHECK: vfcmaddcsh xmm6, xmm5, dword ptr [ecx]
+// CHECK: encoding: [0x62,0xf6,0x57,0x08,0x57,0x31]
+          vfcmaddcsh xmm6, xmm5, dword ptr [ecx]
+
+// CHECK: vfcmaddcsh xmm6, xmm5, dword ptr [ecx + 508]
+// CHECK: encoding: [0x62,0xf6,0x57,0x08,0x57,0x71,0x7f]
+          vfcmaddcsh xmm6, xmm5, dword ptr [ecx + 508]
+
+// CHECK: vfcmaddcsh xmm6 {k7} {z}, xmm5, dword ptr [edx - 512]
+// CHECK: encoding: [0x62,0xf6,0x57,0x8f,0x57,0x72,0x80]
+          vfcmaddcsh xmm6 {k7} {z}, xmm5, dword ptr [edx - 512]
+
+// CHECK: vfcmulcph zmm6, zmm5, zmm4
+// CHECK: encoding: [0x62,0xf6,0x57,0x48,0xd6,0xf4]
+          vfcmulcph zmm6, zmm5, zmm4
+
+// CHECK: vfcmulcph zmm6, zmm5, zmm4, {rn-sae}
+// CHECK: encoding: [0x62,0xf6,0x57,0x18,0xd6,0xf4]
+          vfcmulcph zmm6, zmm5, zmm4, {rn-sae}
+
+// CHECK: vfcmulcph zmm6 {k7}, zmm5, zmmword ptr [esp + 8*esi + 268435456]
+// CHECK: encoding: [0x62,0xf6,0x57,0x4f,0xd6,0xb4,0xf4,0x00,0x00,0x00,0x10]
+          vfcmulcph zmm6 {k7}, zmm5, zmmword ptr [esp + 8*esi + 268435456]
+
+// CHECK: vfcmulcph zmm6, zmm5, dword ptr [ecx]{1to16}
+// CHECK: encoding: [0x62,0xf6,0x57,0x58,0xd6,0x31]
+          vfcmulcph zmm6, zmm5, dword ptr [ecx]{1to16}
+
+// CHECK: vfcmulcph zmm6, zmm5, zmmword ptr [ecx + 8128]
+// CHECK: encoding: [0x62,0xf6,0x57,0x48,0xd6,0x71,0x7f]
+          vfcmulcph zmm6, zmm5, zmmword ptr [ecx + 8128]
+
+// CHECK: vfcmulcph zmm6 {k7} {z}, zmm5, dword ptr [edx - 512]{1to16}
+// CHECK: encoding: [0x62,0xf6,0x57,0xdf,0xd6,0x72,0x80]
+          vfcmulcph zmm6 {k7} {z}, zmm5, dword ptr [edx - 512]{1to16}
+
+// CHECK: vfcmulcsh xmm6, xmm5, xmm4
+// CHECK: encoding: [0x62,0xf6,0x57,0x08,0xd7,0xf4]
+          vfcmulcsh xmm6, xmm5, xmm4
+
+// CHECK: vfcmulcsh xmm6, xmm5, xmm4, {rn-sae}
+// CHECK: encoding: [0x62,0xf6,0x57,0x18,0xd7,0xf4]
+          vfcmulcsh xmm6, xmm5, xmm4, {rn-sae}
+
+// CHECK: vfcmulcsh xmm6 {k7}, xmm5, dword ptr [esp + 8*esi + 268435456]
+// CHECK: encoding: [0x62,0xf6,0x57,0x0f,0xd7,0xb4,0xf4,0x00,0x00,0x00,0x10]
+          vfcmulcsh xmm6 {k7}, xmm5, dword ptr [esp + 8*esi + 268435456]
+
+// CHECK: vfcmulcsh xmm6, xmm5, dword ptr [ecx]
+// CHECK: encoding: [0x62,0xf6,0x57,0x08,0xd7,0x31]
+          vfcmulcsh xmm6, xmm5, dword ptr [ecx]
+
+// CHECK: vfcmulcsh xmm6, xmm5, dword ptr [ecx + 508]
+// CHECK: encoding: [0x62,0xf6,0x57,0x08,0xd7,0x71,0x7f]
+          vfcmulcsh xmm6, xmm5, dword ptr [ecx + 508]
+
+// CHECK: vfcmulcsh xmm6 {k7} {z}, xmm5, dword ptr [edx - 512]
+// CHECK: encoding: [0x62,0xf6,0x57,0x8f,0xd7,0x72,0x80]
+          vfcmulcsh xmm6 {k7} {z}, xmm5, dword ptr [edx - 512]
+
+// CHECK: vfmaddcph zmm6, zmm5, zmm4
+// CHECK: encoding: [0x62,0xf6,0x56,0x48,0x56,0xf4]
+          vfmaddcph zmm6, zmm5, zmm4
+
+// CHECK: vfmaddcph zmm6, zmm5, zmm4, {rn-sae}
+// CHECK: encoding: [0x62,0xf6,0x56,0x18,0x56,0xf4]
+          vfmaddcph zmm6, zmm5, zmm4, {rn-sae}
+
+// CHECK: vfmaddcph zmm6 {k7}, zmm5, zmmword ptr [esp + 8*esi + 268435456]
+// CHECK: encoding: [0x62,0xf6,0x56,0x4f,0x56,0xb4,0xf4,0x00,0x00,0x00,0x10]
+          vfmaddcph zmm6 {k7}, zmm5, zmmword ptr [esp + 8*esi + 268435456]
+
+// CHECK: vfmaddcph zmm6, zmm5, dword ptr [ecx]{1to16}
+// CHECK: encoding: [0x62,0xf6,0x56,0x58,0x56,0x31]
+          vfmaddcph zmm6, zmm5, dword ptr [ecx]{1to16}
+
+// CHECK: vfmaddcph zmm6, zmm5, zmmword ptr [ecx + 8128]
+// CHECK: encoding: [0x62,0xf6,0x56,0x48,0x56,0x71,0x7f]
+          vfmaddcph zmm6, zmm5, zmmword ptr [ecx + 8128]
+
+// CHECK: vfmaddcph zmm6 {k7} {z}, zmm5, dword ptr [edx - 512]{1to16}
+// CHECK: encoding: [0x62,0xf6,0x56,0xdf,0x56,0x72,0x80]
+          vfmaddcph zmm6 {k7} {z}, zmm5, dword ptr [edx - 512]{1to16}
+
+// CHECK: vfmaddcsh xmm6, xmm5, xmm4
+// CHECK: encoding: [0x62,0xf6,0x56,0x08,0x57,0xf4]
+          vfmaddcsh xmm6, xmm5, xmm4
+
+// CHECK: vfmaddcsh xmm6, xmm5, xmm4, {rn-sae}
+// CHECK: encoding: [0x62,0xf6,0x56,0x18,0x57,0xf4]
+          vfmaddcsh xmm6, xmm5, xmm4, {rn-sae}
+
+// CHECK: vfmaddcsh xmm6 {k7}, xmm5, dword ptr [esp + 8*esi + 268435456]
+// CHECK: encoding: [0x62,0xf6,0x56,0x0f,0x57,0xb4,0xf4,0x00,0x00,0x00,0x10]
+          vfmaddcsh xmm6 {k7}, xmm5, dword ptr [esp + 8*esi + 268435456]
+
+// CHECK: vfmaddcsh xmm6, xmm5, dword ptr [ecx]
+// CHECK: encoding: [0x62,0xf6,0x56,0x08,0x57,0x31]
+          vfmaddcsh xmm6, xmm5, dword ptr [ecx]
+
+// CHECK: vfmaddcsh xmm6, xmm5, dword ptr [ecx + 508]
+// CHECK: encoding: [0x62,0xf6,0x56,0x08,0x57,0x71,0x7f]
+          vfmaddcsh xmm6, xmm5, dword ptr [ecx + 508]
+
+// CHECK: vfmaddcsh xmm6 {k7} {z}, xmm5, dword ptr [edx - 512]
+// CHECK: encoding: [0x62,0xf6,0x56,0x8f,0x57,0x72,0x80]
+          vfmaddcsh xmm6 {k7} {z}, xmm5, dword ptr [edx - 512]
+
+// CHECK: vfmulcph zmm6, zmm5, zmm4
+// CHECK: encoding: [0x62,0xf6,0x56,0x48,0xd6,0xf4]
+          vfmulcph zmm6, zmm5, zmm4
+
+// CHECK: vfmulcph zmm6, zmm5, zmm4, {rn-sae}
+// CHECK: encoding: [0x62,0xf6,0x56,0x18,0xd6,0xf4]
+          vfmulcph zmm6, zmm5, zmm4, {rn-sae}
+
+// CHECK: vfmulcph zmm6 {k7}, zmm5, zmmword ptr [esp + 8*esi + 268435456]
+// CHECK: encoding: [0x62,0xf6,0x56,0x4f,0xd6,0xb4,0xf4,0x00,0x00,0x00,0x10]
+          vfmulcph zmm6 {k7}, zmm5, zmmword ptr [esp + 8*esi + 268435456]
+
+// CHECK: vfmulcph zmm6, zmm5, dword ptr [ecx]{1to16}
+// CHECK: encoding: [0x62,0xf6,0x56,0x58,0xd6,0x31]
+          vfmulcph zmm6, zmm5, dword ptr [ecx]{1to16}
+
+// CHECK: vfmulcph zmm6, zmm5, zmmword ptr [ecx + 8128]
+// CHECK: encoding: [0x62,0xf6,0x56,0x48,0xd6,0x71,0x7f]
+          vfmulcph zmm6, zmm5, zmmword ptr [ecx + 8128]
+
+// CHECK: vfmulcph zmm6 {k7} {z}, zmm5, dword ptr [edx - 512]{1to16}
+// CHECK: encoding: [0x62,0xf6,0x56,0xdf,0xd6,0x72,0x80]
+          vfmulcph zmm6 {k7} {z}, zmm5, dword ptr [edx - 512]{1to16}
+
+// CHECK: vfmulcsh xmm6, xmm5, xmm4
+// CHECK: encoding: [0x62,0xf6,0x56,0x08,0xd7,0xf4]
+          vfmulcsh xmm6, xmm5, xmm4
+
+// CHECK: vfmulcsh xmm6, xmm5, xmm4, {rn-sae}
+// CHECK: encoding: [0x62,0xf6,0x56,0x18,0xd7,0xf4]
+          vfmulcsh xmm6, xmm5, xmm4, {rn-sae}
+
+// CHECK: vfmulcsh xmm6 {k7}, xmm5, dword ptr [esp + 8*esi + 268435456]
+// CHECK: encoding: [0x62,0xf6,0x56,0x0f,0xd7,0xb4,0xf4,0x00,0x00,0x00,0x10]
+          vfmulcsh xmm6 {k7}, xmm5, dword ptr [esp + 8*esi + 268435456]
+
+// CHECK: vfmulcsh xmm6, xmm5, dword ptr [ecx]
+// CHECK: encoding: [0x62,0xf6,0x56,0x08,0xd7,0x31]
+          vfmulcsh xmm6, xmm5, dword ptr [ecx]
+
+// CHECK: vfmulcsh xmm6, xmm5, dword ptr [ecx + 508]
+// CHECK: encoding: [0x62,0xf6,0x56,0x08,0xd7,0x71,0x7f]
+          vfmulcsh xmm6, xmm5, dword ptr [ecx + 508]
+
+// CHECK: vfmulcsh xmm6 {k7} {z}, xmm5, dword ptr [edx - 512]
+// CHECK: encoding: [0x62,0xf6,0x56,0x8f,0xd7,0x72,0x80]
+          vfmulcsh xmm6 {k7} {z}, xmm5, dword ptr [edx - 512]

llvm/test/MC/X86/intel-syntax-avx512fp16vl.s

 
 // CHECK: vfnmsub231ph xmm30, xmm29, xmmword ptr [rcx + 2032]
 // CHECK: encoding: [0x62,0x66,0x15,0x00,0xbe,0x71,0x7f]
           vfnmsub231ph xmm30, xmm29, xmmword ptr [rcx + 2032]
 
 // CHECK: vfnmsub231ph xmm30 {k7} {z}, xmm29, word ptr [rdx - 256]{1to8}
 // CHECK: encoding: [0x62,0x66,0x15,0x97,0xbe,0x72,0x80]
           vfnmsub231ph xmm30 {k7} {z}, xmm29, word ptr [rdx - 256]{1to8}
+
+// CHECK: vfcmaddcph ymm30, ymm29, ymm28
+// CHECK: encoding: [0x62,0x06,0x17,0x20,0x56,0xf4]
+          vfcmaddcph ymm30, ymm29, ymm28
+
+// CHECK: vfcmaddcph xmm30, xmm29, xmm28
+// CHECK: encoding: [0x62,0x06,0x17,0x00,0x56,0xf4]
+          vfcmaddcph xmm30, xmm29, xmm28
+
+// CHECK: vfcmaddcph ymm30 {k7}, ymm29, ymmword ptr [rbp + 8*r14 + 268435456]
+// CHECK: encoding: [0x62,0x26,0x17,0x27,0x56,0xb4,0xf5,0x00,0x00,0x00,0x10]
+          vfcmaddcph ymm30 {k7}, ymm29, ymmword ptr [rbp + 8*r14 + 268435456]
+
+// CHECK: vfcmaddcph ymm30, ymm29, dword ptr [r9]{1to8}
+// CHECK: encoding: [0x62,0x46,0x17,0x30,0x56,0x31]
+          vfcmaddcph ymm30, ymm29, dword ptr [r9]{1to8}
+
+// CHECK: vfcmaddcph ymm30, ymm29, ymmword ptr [rcx + 4064]
+// CHECK: encoding: [0x62,0x66,0x17,0x20,0x56,0x71,0x7f]
+          vfcmaddcph ymm30, ymm29, ymmword ptr [rcx + 4064]
+
+// CHECK: vfcmaddcph ymm30 {k7} {z}, ymm29, dword ptr [rdx - 512]{1to8}
+// CHECK: encoding: [0x62,0x66,0x17,0xb7,0x56,0x72,0x80]
+          vfcmaddcph ymm30 {k7} {z}, ymm29, dword ptr [rdx - 512]{1to8}
+
+// CHECK: vfcmaddcph xmm30 {k7}, xmm29, xmmword ptr [rbp + 8*r14 + 268435456]
+// CHECK: encoding: [0x62,0x26,0x17,0x07,0x56,0xb4,0xf5,0x00,0x00,0x00,0x10]
+          vfcmaddcph xmm30 {k7}, xmm29, xmmword ptr [rbp + 8*r14 + 268435456]
+
+// CHECK: vfcmaddcph xmm30, xmm29, dword ptr [r9]{1to4}
+// CHECK: encoding: [0x62,0x46,0x17,0x10,0x56,0x31]
+          vfcmaddcph xmm30, xmm29, dword ptr [r9]{1to4}
+
+// CHECK: vfcmaddcph xmm30, xmm29, xmmword ptr [rcx + 2032]
+// CHECK: encoding: [0x62,0x66,0x17,0x00,0x56,0x71,0x7f]
+          vfcmaddcph xmm30, xmm29, xmmword ptr [rcx + 2032]
+
+// CHECK: vfcmaddcph xmm30 {k7} {z}, xmm29, dword ptr [rdx - 512]{1to4}
+// CHECK: encoding: [0x62,0x66,0x17,0x97,0x56,0x72,0x80]
+          vfcmaddcph xmm30 {k7} {z}, xmm29, dword ptr [rdx - 512]{1to4}
+
+// CHECK: vfcmulcph ymm30, ymm29, ymm28
+// CHECK: encoding: [0x62,0x06,0x17,0x20,0xd6,0xf4]
+          vfcmulcph ymm30, ymm29, ymm28
+
+// CHECK: vfcmulcph xmm30, xmm29, xmm28
+// CHECK: encoding: [0x62,0x06,0x17,0x00,0xd6,0xf4]
+          vfcmulcph xmm30, xmm29, xmm28
+
+// CHECK: vfcmulcph ymm30 {k7}, ymm29, ymmword ptr [rbp + 8*r14 + 268435456]
+// CHECK: encoding: [0x62,0x26,0x17,0x27,0xd6,0xb4,0xf5,0x00,0x00,0x00,0x10]
+          vfcmulcph ymm30 {k7}, ymm29, ymmword ptr [rbp + 8*r14 + 268435456]
+
+// CHECK: vfcmulcph ymm30, ymm29, dword ptr [r9]{1to8}
+// CHECK: encoding: [0x62,0x46,0x17,0x30,0xd6,0x31]
+          vfcmulcph ymm30, ymm29, dword ptr [r9]{1to8}
+
+// CHECK: vfcmulcph ymm30, ymm29, ymmword ptr [rcx + 4064]
+// CHECK: encoding: [0x62,0x66,0x17,0x20,0xd6,0x71,0x7f]
+          vfcmulcph ymm30, ymm29, ymmword ptr [rcx + 4064]
+
+// CHECK: vfcmulcph ymm30 {k7} {z}, ymm29, dword ptr [rdx - 512]{1to8}
+// CHECK: encoding: [0x62,0x66,0x17,0xb7,0xd6,0x72,0x80]
+          vfcmulcph ymm30 {k7} {z}, ymm29, dword ptr [rdx - 512]{1to8}
+
+// CHECK: vfcmulcph xmm30 {k7}, xmm29, xmmword ptr [rbp + 8*r14 + 268435456]
+// CHECK: encoding: [0x62,0x26,0x17,0x07,0xd6,0xb4,0xf5,0x00,0x00,0x00,0x10]
+          vfcmulcph xmm30 {k7}, xmm29, xmmword ptr [rbp + 8*r14 + 268435456]
+
+// CHECK: vfcmulcph xmm30, xmm29, dword ptr [r9]{1to4}
+// CHECK: encoding: [0x62,0x46,0x17,0x10,0xd6,0x31]
+          vfcmulcph xmm30, xmm29, dword ptr [r9]{1to4}
+
+// CHECK: vfcmulcph xmm30, xmm29, xmmword ptr [rcx + 2032]
+// CHECK: encoding: [0x62,0x66,0x17,0x00,0xd6,0x71,0x7f]
+          vfcmulcph xmm30, xmm29, xmmword ptr [rcx + 2032]
+
+// CHECK: vfcmulcph xmm30 {k7} {z}, xmm29, dword ptr [rdx - 512]{1to4}
+// CHECK: encoding: [0x62,0x66,0x17,0x97,0xd6,0x72,0x80]
+          vfcmulcph xmm30 {k7} {z}, xmm29, dword ptr [rdx - 512]{1to4}
+
+// CHECK: vfmaddcph ymm30, ymm29, ymm28
+// CHECK: encoding: [0x62,0x06,0x16,0x20,0x56,0xf4]
+          vfmaddcph ymm30, ymm29, ymm28
+
+// CHECK: vfmaddcph xmm30, xmm29, xmm28
+// CHECK: encoding: [0x62,0x06,0x16,0x00,0x56,0xf4]
+          vfmaddcph xmm30, xmm29, xmm28
+
+// CHECK: vfmaddcph ymm30 {k7}, ymm29, ymmword ptr [rbp + 8*r14 + 268435456]
+// CHECK: encoding: [0x62,0x26,0x16,0x27,0x56,0xb4,0xf5,0x00,0x00,0x00,0x10]
+          vfmaddcph ymm30 {k7}, ymm29, ymmword ptr [rbp + 8*r14 + 268435456]
+
+// CHECK: vfmaddcph ymm30, ymm29, dword ptr [r9]{1to8}
+// CHECK: encoding: [0x62,0x46,0x16,0x30,0x56,0x31]
+          vfmaddcph ymm30, ymm29, dword ptr [r9]{1to8}
+
+// CHECK: vfmaddcph ymm30, ymm29, ymmword ptr [rcx + 4064]
+// CHECK: encoding: [0x62,0x66,0x16,0x20,0x56,0x71,0x7f]
+          vfmaddcph ymm30, ymm29, ymmword ptr [rcx + 4064]
+
+// CHECK: vfmaddcph ymm30 {k7} {z}, ymm29, dword ptr [rdx - 512]{1to8}
+// CHECK: encoding: [0x62,0x66,0x16,0xb7,0x56,0x72,0x80]
+          vfmaddcph ymm30 {k7} {z}, ymm29, dword ptr [rdx - 512]{1to8}
+
+// CHECK: vfmaddcph xmm30 {k7}, xmm29, xmmword ptr [rbp + 8*r14 + 268435456]
+// CHECK: encoding: [0x62,0x26,0x16,0x07,0x56,0xb4,0xf5,0x00,0x00,0x00,0x10]
+          vfmaddcph xmm30 {k7}, xmm29, xmmword ptr [rbp + 8*r14 + 268435456]
+
+// CHECK: vfmaddcph xmm30, xmm29, dword ptr [r9]{1to4}
+// CHECK: encoding: [0x62,0x46,0x16,0x10,0x56,0x31]
+          vfmaddcph xmm30, xmm29, dword ptr [r9]{1to4}
+
+// CHECK: vfmaddcph xmm30, xmm29, xmmword ptr [rcx + 2032]
+// CHECK: encoding: [0x62,0x66,0x16,0x00,0x56,0x71,0x7f]
+          vfmaddcph xmm30, xmm29, xmmword ptr [rcx + 2032]
+
+// CHECK: vfmaddcph xmm30 {k7} {z}, xmm29, dword ptr [rdx - 512]{1to4}
+// CHECK: encoding: [0x62,0x66,0x16,0x97,0x56,0x72,0x80]
+          vfmaddcph xmm30 {k7} {z}, xmm29, dword ptr [rdx - 512]{1to4}
+
+// CHECK: vfmulcph ymm30, ymm29, ymm28
+// CHECK: encoding: [0x62,0x06,0x16,0x20,0xd6,0xf4]
+          vfmulcph ymm30, ymm29, ymm28
+
+// CHECK: vfmulcph xmm30, xmm29, xmm28
+// CHECK: encoding: [0x62,0x06,0x16,0x00,0xd6,0xf4]
+          vfmulcph xmm30, xmm29, xmm28
+
+// CHECK: vfmulcph ymm30 {k7}, ymm29, ymmword ptr [rbp + 8*r14 + 268435456]
+// CHECK: encoding: [0x62,0x26,0x16,0x27,0xd6,0xb4,0xf5,0x00,0x00,0x00,0x10]
+          vfmulcph ymm30 {k7}, ymm29, ymmword ptr [rbp + 8*r14 + 268435456]
+
+// CHECK: vfmulcph ymm30, ymm29, dword ptr [r9]{1to8}
+// CHECK: encoding: [0x62,0x46,0x16,0x30,0xd6,0x31]
+          vfmulcph ymm30, ymm29, dword ptr [r9]{1to8}
+
+// CHECK: vfmulcph ymm30, ymm29, ymmword ptr [rcx + 4064]
+// CHECK: encoding: [0x62,0x66,0x16,0x20,0xd6,0x71,0x7f]
+          vfmulcph ymm30, ymm29, ymmword ptr [rcx + 4064]
+
+// CHECK: vfmulcph ymm30 {k7} {z}, ymm29, dword ptr [rdx - 512]{1to8}
+// CHECK: encoding: [0x62,0x66,0x16,0xb7,0xd6,0x72,0x80]
+          vfmulcph ymm30 {k7} {z}, ymm29, dword ptr [rdx - 512]{1to8}
+
+// CHECK: vfmulcph xmm30 {k7}, xmm29, xmmword ptr [rbp + 8*r14 + 268435456]
+// CHECK: encoding: [0x62,0x26,0x16,0x07,0xd6,0xb4,0xf5,0x00,0x00,0x00,0x10]
+          vfmulcph xmm30 {k7}, xmm29, xmmword ptr [rbp + 8*r14 + 268435456]
+
+// CHECK: vfmulcph xmm30, xmm29, dword ptr [r9]{1to4}
+// CHECK: encoding: [0x62,0x46,0x16,0x10,0xd6,0x31]
+          vfmulcph xmm30, xmm29, dword ptr [r9]{1to4}
+
+// CHECK: vfmulcph xmm30, xmm29, xmmword ptr [rcx + 2032]
+// CHECK: encoding: [0x62,0x66,0x16,0x00,0xd6,0x71,0x7f]
+          vfmulcph xmm30, xmm29, xmmword ptr [rcx + 2032]
+
+// CHECK: vfmulcph xmm30 {k7} {z}, xmm29, dword ptr [rdx - 512]{1to4}
+// CHECK: encoding: [0x62,0x66,0x16,0x97,0xd6,0x72,0x80]
+          vfmulcph xmm30 {k7} {z}, xmm29, dword ptr [rdx - 512]{1to4}